domestic alliance network to attract foreign partners: evidence from

Domestic alliance network to attract foreignpartners: Evidence from international jointventures in China

Weilei (Stone) Shi1, Sunny LiSun2, Brian C Pinkham3 andMike W Peng4

1Zicklin School of Business, Baruch College, CityUniversity of New York, New York, USA; 2Institutefor Entrepreneurship and Innovation, Bloch Schoolof Business and Administration, University ofMissouri, Kansas City, USA; 3Ivey Business School,Western University, London, Canada; 4School ofManagement, University of Texas at Dallas,Richardson, USA

Correspondence:W (S) Shi, Zicklin School of Business, BaruchCollege, City University of New York, VerticalCampus 9-281, One Bernard Baruch Way,New York, NY 10010, USA.Tel: (646) 312 3660;Fax: (646) 312 3621;email: [email protected]

Received: 10 February 2012Revised: 12 November 2013Accepted: 5 December 2013Online publication date: 23 January 2014

AbstractPartner selection is a critical issue in building international joint ventures (IJVs).We argue that foreign firms are more likely to select local firms with uniquenetwork structural advantages within a local alliance network. We framestructural advantages as two network position traits: centrality and brokerage.Specifically, network centrality acts as a stronger network trait than brokerage inattracting foreign IJV partners. However, such a relationship may be moderatedby foreign firms’ local experience and perceived capabilities. We contend thatwhen foreign firms have a high level of local market experience and perceivedcapabilities, they may prefer a local broker over a centrally located local firm.Data on the domestic alliance network in China’s electronics and informationtechnology (IT) industries largely support our hypotheses. We conclude that asforeign investors become strategic insiders, they may not only seek a localpartner’s capability attributes, but also more critically pay attention to a localpartner’s domestic network.Journal of International Business Studies (2014) 45, 338–362. doi:10.1057/jibs.2013.71

Keywords: alliance network; partner selection; alliances and joint ventures; internationaljoint ventures (IJVs); centrality; China

INTRODUCTIONHow can foreign firms identify the right local firms as internationaljoint venture (IJV) partners? Research on partner selection hasfocused on partner-related characteristics, such as partner attractive-ness, complementarity, and alliance experience (Hitt, Ahlstrom,Dacin, Levitas, & Svobodina, 2004; Li, Zhou, & Zajac, 2009; Luo,1997, 1998, 2008; Mitsuhashi & Greve, 2009; Zhao, Anand, &Mitchell, 2005). Within the alliance formation literature, somestudies focus on dyad-level constructs such as tie strength (Wong &Ellis, 2002). Other studies deal with features of the overall alliancenetwork, such as the clustering patterns among partners (Nair,Hanvanich, & Cavusgil, 2007) and the multiplexity of the network(Beckman, Haunschild, & Phillips, 2004). The existing IJV literature,however, has paid less attention to local partners’ network positionsas an important trait for partner selection, despite the critical rolesof network positions in firms’ strategy and competitive advantagedocumented by Burt (1992) and Koka and Prescott (2008). It has also

Journal of International Business Studies (2014) 45, 338–362© 2014 Academy of International Business All rights reserved 0047-2506

www.jibs.net

mailto:[email protected]

http://dx.doi.org/10.1057/jibs.2013.71

http://www.jibs.net

paid inadequate attention to the actual practice ofdue diligence process in IJV partner selection inwhich multinational enterprises (MNEs) often inves-tigate focal firms’ network position to comprehen-sively map the local embeddedness (Wolf, 2000). Inthis article, we focus on how individual local firms’network position traits – primarily centrality andbrokerage – within the domestic alliance networkinfluence foreign firms’ partner selection.We define a “centrally located local firm” as a

firm that occupies a central position in a domesticalliance network, and a “local broker” as a firm thatbridges a structural hole in a domestic alliance net-work (Yang, Lin, & Peng, 2011a). A firm’s networkpositions have a profound impact on its partneringbehavior (Benjamin & Podolny, 1999; Podolny,1993, 2001). Some research emphasizes centrality(Higgins & Gulati, 2003; Ozmel, Reuer, & Gulati,2012; Podolny, 2001), but does not examine broker-age. Other research has shown that brokerage isimportant for partner selection (Guler & Guillén,2010; Jensen, 2008; Xiao & Tsui, 2007). For example,Xiao and Tsui (2007: 5) argue that firms bridgingstructural holes may be unattractive partners because“people who stay at the boundary of two in-groupstend to be distrusted by both groups”. This streamassumes that brokerage is observable and may thusinfluence partner selection.Extending such earlier work, we argue that a firm’s

network positions in the domestic alliance network –

measured by centrality and brokerage – influence itsodds of being selected as IJV partners by foreignentrants. Some previous studies suggest that net-work centrality matters for partner selection, but fallshort of discussing under what conditions centralitymatters more relative to brokerage. We endeavor toshed new light on this line of research by investigat-ing how centrally located local firms differ from localbrokers in their odds of being selected as IJV part-ners. Such a comparison is important since priorstudies have not differentiated between the relativebenefits of each network position trait (Pollock &Gulati, 2007). In the context of market entry into anemerging economy (Meyer, Estrin, Bhaumik, &Peng, 2009a), we also study the boundary conditionsof network position traits as selection criteria in theprocess of IJV partner selection.Specifically, we focus on two questions: (1) How

do foreign firms observe and distinguish betweencentrally located local firms and local brokers inthe local alliance network during partner selection?(2) How do foreign firms’ local experience andperceived capabilities influence their choice between

centrally located local firms and local brokers asIJV partners? Our core argument is that networkposition traits differ in their visibility and level ofinimitability. We hypothesize that the effect of acentrally located local firm on partner selection willbe stronger than that of a local broker. We also arguethat foreign firms’ local experience and perceivedcapabilities can help them identify these networkposition traits. Where traditional IJV partner selec-tion research suggests that foreign firms with highcapabilities may seek out centrally located local firms(Benjamin & Podolny, 1999), we argue that foreignfirms may consider local brokers and centrallylocated local firms in a contingent manner. In otherwords, we argue that when foreign firms have alonger presence in the host country and when theyare more capable, they are more likely to select alocal broker than a centrally located local firm.Figure 1 illustrates our research framework.This article aspires to make two contributions.

First, building on social network theory, we contri-bute to the literature on IJV partner selection bydemonstrating how network position traits mayattract foreign partners in economies where moralhazards and adverse selection have been key pro-blems for foreign firms. Unlike prior studies thatfocus mainly on centrally located local firms, weexamine both centrally located local firms and localbrokers under the assumption that both networkposition traits can be observed, but the level ofeffort needed to observe and interpret these positiontraits in a local network differ widely. Second,we also enrich our understanding of the comparativevalue of both network position traits in emerging

DomesticAllianceNetwork

Foreign Firm’s CapabilitiesPerceived by Local Firms

(Fortune 500 Membership)

CentrallyLocatedLocal Firm

LocalBroker

PartnerSelection

Foreign Firm’s Local MarketExperience

H1a (+)

H1b (+)

H3b (+)H3a (+)

H4b (+)H4a (+)

Figure 1 Theoretical framework.Note: Hypotheses 2, 3c, and 4c are not shown because theycompare the relative effects of network centrality and brokerage.

Domestic alliance network to attract foreign partners Weilei (Stone) Shi et al339

Journal of International Business Studies

economies. Existing research tends to study whenand under which condition network position traitsmatter individually. Our study aims to understandunder which conditions one type of network posi-tion trait (such as being a local broker) matters morethan the other type (such as being a centrally locatedlocal firm). Overall, our comparative contingencyapproach helps establish boundary conditions forIJV partner selection in emerging economies.

THEORETICAL DEVELOPMENT

Importance of Local Network and Its PositionWe argue that local network and its position inemerging economies are becoming increasingly impor-tant for foreign firms for three reasons. First, manyforeign investors have been transformed into stra-tegic insiders that are significantly building andlocalizing their primary activities of the entire valuechain in these markets (Luo, 2007). Such strategicinsiders particularly rely on local networks, espe-cially those local firms holding central or brokeragepositions. Second, due to heightened industriali-zation and specialization in emerging economies,businesses are becoming increasingly interdepen-dent, particularly in resource sharing and synergeticcollaboration, which makes local networks and posi-tions within such networks more critical than everbefore (Zhao et al., 2005). In this environment,understanding the embedded local norms and prac-tices is of utmost importance (Laursen, Masciarelli, &Prencipe, 2012). Third, historically and culturally,success in China depends not merely on how exten-sive a firm’s networks are but more on with whom itnetworks (Peng & Luo, 2000). This is particularlytrue considering that the institutional environmentis still underdeveloped, characterized by informationasymmetries and critical resources still being con-trolled by governmental institutions (Krug &Hendrischke, 2012; Li, Peng, & Macaulay, 2013; Shi,Markóczy, & Stan, 2013). Thus local partners hold-ing central or brokerage positions will be potentiallymore valuable to foreign firms.

Network Position Traits in an Emerging EconomyWhile prior research on IJV partner selection hasfocused on a variety of selection criteria, two net-work position traits in particular can reveal impor-tant information about a local firm’s resources andprospects in emerging economies. First, IJV scholarshave argued that identifying specific selection cri-teria may depend on the strategic context (Geringer,1991; Luo, 1997). In an emerging economy context,

accessing local cultural, regulatory, and marketknowledge are the three top concerns of foreignfirms (Meyer et al., 2009a; Tatoglu & Glaister, 2000).Prior network research has well documented the roleof brokerage in accessing such information (Guler &Guillén, 2010; Koka & Prescott, 2008). In China,recent studies indicate that the local network repre-sents an alternative channel for critical resources andinformation to flow among market players dueto the significant advancement and developmentof various business networks (Keister, 2009). Thissituation is particularly salient in Chinese high-technology industries (our research setting) – sug-gesting the need to access information throughvarious business networks (Li & Zhang, 2007).Meanwhile, recent IJV research indicates that

under the condition of underdeveloped regulatoryregime where the problem of adverse selectionabounds (Roy & Oliver, 2009), foreign firms tendto focus on factors such as local firms’ industryposition, prominence within the local network, andpartnering ability (Damanpour, Devece, Chen, &Pothukuchi, 2012). As noted earlier, local firms’ net-work centrality is clearly related to these importantfactors. Therefore given the context in which westudy, our focus on the network position as an IJVselection criterion is not only relevant but alsoimportant (Glaister & Buckley, 1997). Specifically,the China context provides a unique opportunityto assess the impact of local firms’ centrality andbrokerage advantages.Second, local partners’ network position is a strate-

gic attribute that remains under-researched in partnerselection research in emerging economies and parti-cularly in China (Yang, Sun, Lin, & Peng, 2011b).Recent research shows that success in China increa-singly depends on the cultivation of ego-businessnetworks such as vertical networks (Liu, Luo, & Liu,2009) and horizontal networks (Yang et al., 2011a).In addition, equity-based joint ventures or alliancesare recently used not only to access local partners’distinctive resources and capabilities, but moreimportantly to access their network-based resourcesand capabilities (Lin, Peng, Yang, & Sun, 2009a). Assuch, network position traits such as centrality andbrokerage now matter more to foreign firms’ growthand success in China. However, an analysis of theimportance of brokerage and centrality to foreigninvestors has not been addressed in the IJV literature.It is, therefore, both theoretically and empiricallyimperative that this factor be investigated.Despite attractive market potential in an emer-

ging economy, foreign firms often face significant



uncertainties (Meyer et al., 2009a) and liabilities offoreignness (Kostova & Zaheer, 1999). The competi-tive space in an emerging economy, far from being“perfectly competitive,” is characterized by imper-fect and incomplete information (Jacobson, 1992;Peng, 2003). The literature has documented twodifferent sources of such liabilities of foreignness.The first is the lack of information and knowledgeabout local market in terms of consumer behavior,culture, and legal environment (Guler & Guillén,2010). The second is the lack of legitimacy in thenew market (Zaheer, 1995) and the informationabout local partners’ partnering ability (Roy &Oliver, 2009). We argue that local firms that arewell connected in the domestic alliance networkcan help foreign firms to overcome some of theseliabilities (Lu & Ma, 2008) and to establish legiti-mization and develop partnering competence(Dacin, Oliver, & Roy, 2007).Following Burt (1992), we focus on centrally

located local firms (centrality) and local broker firms(brokerage) for two reasons. First, these two net-work position traits have been documented in priorstudies in emerging economies to be importantand relevant (Lin et al., 2009a). Second, within net-work studies, different levels of analyses may allowresearchers to study different network features andrelated questions. Our main interest is how indivi-dual firms’ network features within the overall net-work impact partner selection behaviors. Our choiceof two network position traits is particularly inter-esting in the setting of IJV formation in an emer-ging economy where foreign firms are usually lessembedded within the domestic network thandomestic firms. Such asymmetric levels of embedd-edness (Ahuja, Polidoro, & Mitchell, 2009) requireus to fully explore the incentives and constraintsfor creating ties between local firms and foreignpartners.Overall, we argue that centrally located local firms

and local brokers convey different resources andinformation to foreign firms. While our baselinemodel predicts that both centrally located local firmsand local brokers are likely to be selected by foreignpartners as IJV partners, our comparative modelsuggests that centrally located local firms in generalare more likely to be selected as potential partners.

HYPOTHESIS DEVELOPMENT

Firm Centrality Position TraitFirms that occupy a central position in a socialnetwork generally enjoy a high level of social status

(Podolny, 1993). Social status conveys importantinformation about a firm’s underlying quality(Jensen, 2008). Within the alliance context, a firm’scentral position trait represents its resources andprospects, which alleviates problems associated withadverse selection (Shipilov & Li, 2008). “The factthat a particular firm is able to form extensive directand indirect ties in alliance networks and henceachieve a prominent position indicates that thefirm possesses valuable resource and capabilities thatare in demand by other companies” (Ozmel et al.,2012: 10).Within our research context, we argue that cen-

trally located local firms have two underlying capa-bilities that are important for IJV formation withforeign partners. The first is the ability of local firmsto establish legitimacy and gain prominence inthe local market. Centrally located local firms enjoydeference from others due to their standing in thesocial hierarchy (Gould, 2002; Guler & Guillén,2010). Foreign firms are attracted to potential part-nerships with centrally located local firms becausethe high social status of these local firms canconfer legitimacy – an important, intangibleresource that may help foreign firms overcome theinitial liabilities of foreignness (Kostova & Zaheer,1999). In addition, centrality also indicates marketpower, which is widely sought by foreign partners.This is because this power fortifies a foreign inves-tor’s ability to cope with market uncertaintiesand institutional hindrance. For example, Lin et al.(2009a) argue that centrally located firms in Chinahave a higher likelihood of establishing a footholdand leadership within the local marketplace. Thismarket power can usually translate into greaterability to access partners’ resources (Okhmatovskiy,2010) and higher bargaining power with externalstakeholders (Luo, 1997). Ren, Au, and Birtch (2009)find that centrally located firms in China tend toreceive a significant amount of loan guarantees fromtheir partners. Markóczy, Sun, Peng, Shi, and Ren(2013) also point out that a Chinese firm occupyinga central position tends to have stronger cooptationforces that mitigate the threat of potential stateintervention, which is a critical capability that ishighly sought after by foreign partners given theinstitutional uncertainty and business risk in China(Luo, 1997).The second is the partnering capability. When

facing information asymmetries in an emergingeconomy, choosing the wrong local firms maygenerate a significant headache (Ding, Huang, &Liu, 2012). Social status positively influences the



collective perceptions of potential partners in regardto the focal firm’s quality and ability to deliver on itscommitments in the future (Guler & Guillén, 2010;Pollock & Gulati, 2007), particularly when informa-tion about the quality and behavior is imperfectlyobservable (Podolny, 2001). Therefore the fact thatthe focal firm has multiple linkages with differentlocal firms indicates that it possesses sufficient skillsin developing high-quality relationships with otherfirms and is a trustworthy partner (Roy & Oliver,2009).

Hypothesis 1a: A centrally located local firm’sposition (centrality) within the domestic alliancenetwork is positively related to its likelihood ofbeing selected as an IJV partner by foreign firms.

Local Broker Position TraitLocal brokers span different segments within thelocal market. Their unique position allows them toaccess non-redundant (usually novel) opportuni-ties and localized information (Burt, 2007).We arguethat foreign firms are particularly eager to acquiresuch local knowledge in an effort to overcome liabi-lities of foreignness due to their lack of local infor-mation (Guler & Guillén, 2010; Meyer, Wright, &Pruthi, 2009b). In particular, the local market, cul-tural, and regulatory knowledge appears to be thetop three concerns for foreign firms entering emer-ging economies (Nielsen, 2003). Local marketknowledge refers to market-specific knowledge suchas information on market forecasts, consumer pre-ferences, and supplier behaviors (Figueiredo & Brito,2011; Mariotti & Piscitello, 1995). Local cultural andregulatory knowledge refers to information regard-ing local institutions such as local procedures andpractices of law (Glaister & Buckley, 1997; Nielsen,2003). Such knowledge is likely to be institutionallysensitive and embedded within the local con-text (Tan & Meyer, 2011). It is socially constructed,tacit, and idiosyncratic in nature (Dhanaraj, Lyles,Steensma, & Tihanyi, 2004).By definition, the advantage accrued by local

brokers is usually local (Burt, 2007; Shi, Markóczy,& Dess, 2009) and context-specific (Guler & Guillén,2010; Koka & Prescott, 2008). Such advantage maylose its strategic value out of the local firms’ imme-diate network. Local brokers generally connect withfewer local firms than centrally located local firms.It is, therefore, easy for local brokers to spend moretime to interact and socialize with each specific firmin the local context. This helps local brokers to gaina rich understanding of the highly embedded and

sensitive local information (Laursen et al., 2012).The nature of local embeddedness enables theselocal brokers to better understand the needs oflocal consumers that have been traditionallyignored, leading to potential new opportunities forsales and profitability. Local brokers may also havea better position to understand local rules of thumb,regulation, and possibly create value by bridginginstitutional distance (Tracey & Phillips, 2011) oreven institutional contradictions (Bjerregaard &Lauring, 2012) within a single emerging economy(Chabowski, Hult, Kiyak, & Mena, 2010; Meyer &Nguyen, 2005).In fact, local brokers are particularly important in

China since information transmission is quite stickydue to a lack of nationwide markets and inefficientmarket intermediaries (Chang & Xu, 2008). Localbrokers therefore can create a competitive advantageby overcoming the geographic limitations of infor-mation transfer (Owen-Smith & Powell, 2004),thereby tapping into more diverse, non-redundant,and novel knowledge (Winter, 2012).However, how can a foreign firm identify a local

broker? We suggest that there are both “pull” and“push” arguments underlying the local broker iden-tification process. Specifically, local brokers oftenshare non-overlapping information with alters thataid in opportunity recognition (Koka & Prescott,2008). Brokers are also often necessary to matchunconnected parties in order to realize a specifictransaction. The need for local brokers’ services incen-tivizes brokers to “advertise” themselves (Markóczyet al., 2013). This is one of the ways why foreignfirms may identify specific local broker firms(Galunic, Ertug, & Gargiulo, 2012).Furthermore, recent studies also suggest that firms

that bridge structural holes also tend to exhibitbehaviors that make it easy to discern to theirpartners (Sytch, Tatarynowicz, & Gulati, 2012). Ina network of interorganizational partnerships in thecomputer industry, for example, brokers are oftenkey hubs for recombinant innovations, whichendow them with significant power relative to pro-viders of a single or substitute components (non-brokers). Non-brokers depend on the brokers’ abilityto supply new ideas for recombinant innovationsand complementary inputs (Ding et al., 2012). Thistypically translates into aggressive contract negotia-tion practices by the brokers with unevenly distrib-uted spoils. This behavior (as a byproduct ofbrokerage) can affect the brokers’ partners directly,and is thus relatively easy for these firms to actupon (Sytch et al., 2012). Our additional open-ended



interviews with two top managers in China aboutthe discernibility of dealing with brokers reinforcethese points. The interviewees suggested that whilelocal brokers tend to boast about their abilities toprovide access to other alters, they also tend to care-fully guard direct access to these alters. The inter-viewees gave other specific examples for discerniblebehaviors that characterize local brokers. In informalsettings, for example, local brokers are extremelyeager to gather information about other firms butare reluctant to share information about themselvesbeyond emphasizing that they have access to altersfrom which the interested partners would benefit.Local brokers also tend to seek invitations to infor-mal social gatherings, events, and parties that areseemingly irrelevant to their domain of business.These arguments suggest:

Hypothesis 1b: A local broker’s position (broker-age) within the domestic alliance network is posi-tively related to its likelihood of being selected asan IJV partner by foreign firms.

Centrality vs Brokerage: A ComparisonOur arguments on centrally located local firms andlocal brokers indicate that both types of firms arepositively related to the likelihood of being selectedas IJV partners by foreign firms, although eachprovides different resources. However, which net-work position traits have a stronger likelihood inattracting foreign firms? To address this question,we argue network position traits differ by (1) theirvisibility and (2) their degree of inimitability. Thebrokerage trait, compared with the centrality trait, isless visible and therefore more difficult to identify.Meanwhile it is also more difficult to imitate acentrally located local firm than a local broker.A firm’s proclivity to form alliances is largely

dependent on its visibility within the industry(Pollock, Rindova, &Maggitti, 2008). In other words,foreign partners need to first be aware of a given localfirm’s existence and availability in order to form apotential alliance with it (Pollock & Gulati, 2007).For foreign firms, a centrally located local firm isobservable prior to the actual formation of thepotential IJV (Al-Laham & Souitaris, 2008). This isbecause information on local firms’ previous alli-ances within a domestic alliance network is availablethrough at least three credible sources. First, forlisted firms, such information is typically includedin the registration statement to the relevant regula-tory authorities, such as the China Securities Regula-tory Commission (CSRC),1 which is similar to the

Securities and Exchange Commission in the UnitedStates. Second, for both listed and non-listed localfirms, such information may be available on a localfirm’s website or through initial contact or inter-views. Third, foreign firms can triangulate the localfirm’s network position from media sources.2 Themedia in an emerging economy may be especiallyinterested in reporting alliance-related activities,which create jobs and boost local prosperity. Mediaexposure of firms’ alliance activities tends to increasethe chance that a focal firm will be noticed byforeign partners. In fact, stakeholders often rely onmedia sources as important disseminators of infor-mation about firms, particularly when uncertainty ishigh (Zavyalova, Pfarrer, Reger, & Shapiro, 2012). Inthis way, we follow Podolny’s (1993) argument thathigh status (through its premier position within analliance network) creates certain benefits for thelocal firm. Media attention may supplement a firm’sown advertisement and may improve exposure topotential recipients. Furthermore, this exposure alsohelps foreign partners to perceive local firms’ net-work position traits in a positive way.Meanwhile, a high degree of centrality in the

domestic alliance network is difficult and costly toimitate by other local firms because a high centralitynetwork position is not built overnight (Koka &Prescott, 2008). A firm must make large initialinvestments in forming and managing the alliancenetwork (Lavie, 2007). Therefore only a limitednumber of local firms can successfully positionthemselves to be centrally located within a domesticalliance network. In this sense, the costs of position-ing a firm in a central location in the local networkcreate exclusivity among the small number of localfirms that may occupy such a position (Uzzi &Lancaster, 2005). In turn, the network positionexclusivity makes it costly (and difficult) for lesscentrally located local firms to mislead foreign firmsseeking centrally located IJV partners. Similarly,broker firms need to be particularly sensitive to theoverall network changes and knowwho is connectedor not connected with whom. The social process toestablish and maintain these connections createscausal ambiguity and complexity. The competenceunderlying this type of activity is established overtime, making it costly for pretenders to imitate in theshort-run.Compared with locating centrally located firms, it

is more challenging for foreign firms to identify localbrokers for two reasons. First, Burt (2005: 22–23)argues that “seeing” brokerage positions is not easy.To accurately identify the local brokers, a foreign



firm must accurately map both the local firm’snetwork and that particular local firm’s partners’network(s). The amount of information required toaccurately map such complex networks becomesexponentially demanding (Krackhardt, 1987). Second,social networks evolve and change over time (Koka& Prescott, 2008). Local brokers’ network positionsmay be especially unstable compared with centrallylocated local firms’ positions (Lin et al., 2009a). Forinstance, firms associated with the local broker mayform ties with each other in an attempt to blockvalue appropriation by the local broker (Madhavan,Gnyawali, & He, 2004). Alternatively, local brokersmay develop ties with distant firms in an effortto pool resources, making it difficult for outsideobservers to accurately reconstruct the local network(Obstfeld, 2005; Shi et al., 2009). Such instabilitymay reduce the visibility of local brokers by foreignfirms. Overall, these arguments suggest that a localbroker may be less visible to foreign partners than acentrally located local firm. The level of visibility,as prior literature shows, may influence the like-lihood of inclusion in the “consideration set”(Roberts & Lattin, 1997) of organizations that could,for instance, become valuable alliance partners(Pollock & Gulati, 2007).Finally, the value of local brokers is in access to

tacit local knowledge at low cost. These firms preferweak ties over strong ties (Burt, 1992). Initiatinga network of mainly weak ties is less costly thanestablishing a network of strong ties typified by acentrally located local firm (Levin & Cross, 2004).Therefore it may be comparatively easier to imitatelocal brokers than to imitate centrally located localfirms.In summary, identifying local brokers is difficult,

but not impossible. However, foreign firms mustwork harder to identify these local brokers. In otherwords, it is easier for foreign firms to observe andidentify centrally located local firms than local bro-kers. These arguments suggest the following:

Hypothesis 2: The relationship between a localbroker’s position (brokerage) within the domesticalliance network and its likelihood of being selec-ted as an IJV partner by foreign firms will beweaker than the relationship between a centrallylocated local firm’s position (centrality) and thelikelihood of being selected by foreign firms.

Contingency Effect of Network Position TraitsVisibility is important for foreign firms to identifycentrally located local firms and local brokers.

Underlying this statement are two importantassumptions. First, foreign firms have the sameability to accurately observe these network positiontraits. Second, local firms willingly demonstrateand therefore will not purposefully mask their net-work position traits. In developing our arguments onmoderating effects, we now relax these two assump-tions. The ability to identify network traits differsamong foreign firms. In particular, firm-specific cap-abilities can influence the interpretation and trans-mission process of network traits (Heil & Robertson,1991). In addition, local firms vary in their will-ingness to demonstrate their network position traits.Therefore it is important to investigate (1) foreignfirms’ local market experience and (2) foreign firms’capabilities as perceived by local firms.

Foreign firms’ local market experienceOne important moderating variable is organiza-tional experience – the intensity of exposure tocertain activities or the time spent on these activities(Luo & Peng, 1999). We focus on a foreign firm’slocal market experience because it indicates famil-iarity with the local market. In addition, local marketexperience may capture the foreign firm’s identifica-tion processing capabilities (Heil & Robertson,1991). The foreign firm’s capabilities to identify localfirms’ network position traits is critical during part-ner selection, allowing the foreign firm to accuratelyidentify centrally located local firms and local bro-kers (Luo & Peng, 1999).Local market experience may thus help the foreign

firm in at least two ways. First, it improves theefficiency of information search and processing(Chang, Gong, & Peng, 2012). Second, foreign firmswith a high level of local experience may have well-developed cognitive structures capable of piecingtogether a local firm’s network portfolio (Ozcan &Eisenhardt, 2009). Foreign firms with a high level oflocal experience are also more likely to enjoy goodrelationships locally, thus facilitating the informa-tion search process (Li, Poppo, & Zhou, 2008).

Hypothesis 3a: A foreign firm’s experience ina local market will positively moderate the rela-tionship between centrally located local firms’positions (centrality) within the domestic alliancenetwork and the likelihood of being selected as IJVpartners by foreign firms.

Hypothesis 3b: A foreign firm’s experience in alocal market will positively moderate the relation-ship between local brokers’ positions (brokerage)within the domestic alliance network and the



likelihood of being selected as IJV partners byforeign firms.

The above arguments suggest that foreign firms’local market experience helps them understandnetwork position traits (regardless of centrality orbrokerage) more effectively. However, correspondingto Hypothesis 2, what is the role of foreign firms’ localmarket experience in their selection of centrallylocated local firms relative to local brokers? Next, weargue that when foreign firms’ local market experi-ence is strong, they are more likely to select localbrokers rather than centrally located local firms.Firms with different levels of local market experi-

ence may have different challenges and thereforeexperience different liabilities of foreignness (Luo,2007). First, for foreign firms with a low level of localmarket experience, achieving local legitimacy thro-ugh building technological leadership or definingthe technological standard in the local market maybe a primary concern (Isobe, Makino, &Montgemery,2000; Luo & Tung, 2007). This may provide certainadvantages in an emerging economy where estab-lished technical specifications and designs amonglocal firms rarely exist (Yan, 1998). The main goal forthese foreign firms is to survive rather than to makeprofit. Foreign firms with little local experience maytherefore like to draw on centrally located local firmsto achieve this goal of gaining local legitimacy.In contrast, foreign firms with a high level of local

experience have gained initial local legitimacy andare more likely to be concerned with expanding tonew territories (either geographically or product-related) within an emerging economy. To thesefirms, the change of concern and challenge makescentrally located local firms less attractive. In con-trast, local brokers may become more attractive IJVpartners because they provide complementary,novel information that foreign firms can use formarket expansion.Foreign firms with a high level of local experience

may want to explore the idiosyncratic local rules andnorms in various regions or areas. Similar to tourists,foreign firms may seek a mainstream tour guideto show them the high-profile sights on an initialtrip to the local market. However, on subsequenttrips, tourists, having seen the major sights, mayseek a more specialized tour guide that specializes in“off the beaten path” sights and experiences in thelocal market. From this perspective, we may considerthe local network a field of secondary brokers thatpossess information about local interests that mayexclude high-profile sights all together, and thus are

more interesting to a seasoned traveler – or, in ourcase, an experienced foreign firm (Burt, 2007). There-fore local market, cultural, and regulatory knowledgeand information become increasingly importantand arise as the top priority for experienced foreignfirms.Meanwhile, firms with little local experience may

also suffer from information asymmetries regardingtheir potential local partners’ trustworthiness (Guler& Guillén, 2010; Luo & Peng, 1999; Roy & Oliver,2009). Experience in the local market helps foreignfirms to cultivate important organizational routinesthat mitigate adverse selection (Ozmel et al., 2012).Simply put, experienced foreign firms tend to recog-nize potentially trustworthy local partners. There-fore selecting centrally located local firms becomesless important as a source of identifying trustworthylocal partners.In addition, foreign firms with a high level of local

experience may be more concerned with technologyspillovers that may increase rivalry (Meyer & Sinani,2009). Centrally located local firms are in the bestposition to take advantage of technology spillovers.In contrast, local brokers specialize in niche mar-kets (either technologically or geographically), andthus are less likely to have technical domains thatoverlap with the foreign firm. Limited absorptivecapability will also keep local brokers from takingadvantage of technology spillovers from foreignfirms. The breadth of technological coverage by acentrally located local firm may limit the absorptionof the technology spillovers from foreign firms.However, centrally located local firms may effi-ciently diffuse technology know-how to other localfirms given their local alliance network position.3

Taken together:

Hypothesis 3c: The moderating effect of foreignfirms’ experience in a local market will be strongerin the selection of local brokers than in the selec-tion of centrally located local firms. In otherwords, experienced foreign firms are more likelyto select local brokers than centrally located localfirms within the local alliance network.

Foreign firms’ capabilities as perceived by local firmsAs discussed earlier, local partners have incentives toadvertise their network positions. However, thedesire to do so varies by conditions. One of theimportant conditions is foreign firms’ perceivedcapabilities, which is defined as the capabilities thatare perceived by local firms. When foreign firms’perceived capabilities are strong, local partners’



desire to demonstrate network position traits will behigher. In other words, the utility of demonstratingnetwork position traits is essentially a function oflocal firms’ willingness to enhance their visibility inthe eyes of foreign firms (Higgins & Gulati, 2003;Pollock et al., 2008). This is because a fundamentalrequirement for membership in local networks is theperceived ability to reciprocate in exchange relations(Choi & Beamish, 2013). Trading information is partof an exchange relationship that is bounded bythe rules of reciprocity (Puffer, McCarthy, & Peng,2013; Teagarden & Schotter, 2013; Thams, Liu, &Von Glinow, 2013). In addition, a trader may “favorpartners that promise themost useful information inreturn for his/her own” (Carter, 1989: 159). Theability of a foreign firm to contribute valuable infor-mation and skill-sets in a potential partnership withlocal firms will depend on the amount of resourcesthat it is perceived (by local firms) to possess (Luo,2008; Zhan, Chen, Erramilli, & Nguyen, 2009). Tothe same extent that foreign firms prefer morecapable local partners, local firms also prefer morecapable foreign firms (Mitsuhashi & Greve, 2009;Zhan & Chen, 2013). We argue that a more capableforeign firm (in the eyes of local partners) will bemore attractive to local firms. In turn, the perceivedcapabilities will help a foreign firm accelerate itslearning of the local network, enhancing its abilityto interpret network position traits. Specifically,

Hypothesis 4a: A foreign firm’s perceived cap-abilities in the eyes of local firms will positivelymoderate the relationship between local brokers’positions (brokerage) within the domestic alliancenetwork and the likelihood of being selected as IJVpartners by foreign firms.

Hypothesis 4b: A foreign firm’s perceived cap-abilities in the eyes of local firms will positivelymoderate the relationship between centrally loca-ted local firms’ positions (centrality) within thedomestic alliance network and the likelihood ofbeing selected as IJV partners by foreign firms.

Here we are also interested in examining therelative importance of centrally located local firmsvs local brokers under the condition of foreign firms’capabilities. In parallel with Hypothesis 3c, we arguewhen foreign firms’ capabilities are perceived higherby local firms, foreign firms are more likely toselect local brokers rather than centrally locatedlocal firms. This is because foreign firms with a highlevel of perceived capabilities (such as belonging toFortune 500) already possess a high status globally

and command a great deal of respect locally. Thisis especially true in an emerging economy suchas China where foreign products (especially thoseproduced by reputable foreign firms) are usuallyperceived as higher quality than domestic ones(Brouthers & Xu, 2002). This indicates that foreignfirms that have a high level of perceived capabilitiescan achieve local legitimacy even without ties withcentrally located local firms. Therefore these foreignfirms may be less likely to ally with centrally locatedlocal firms given that these local firms’ statusresources may be redundant (Lin et al., 2009a). Cen-trally located local firms may also have extravagantdemands on the foreign firms, increasing the costof IJVs for foreign firms (Tong, Reuer, & Peng, 2008).In addition, the overlapping reputation resourcesmay push centrally located local firms into a compe-titive position with these foreign firms.In contrast, local brokers usually possesses comple-

mentary (but non-overlapping) knowledge such asunique local knowledge and region-specific expertise(Shipilov & Li, 2008). In this sense, foreign firms witha high level of perceived capabilities may find a localbroker more attractive as an IJV partner relative toa centrally located local firm. Furthermore, foreignfirms with a high level of perceived capabilities mayhave a much larger pool of resources to expand with-in an emerging economy. Understanding the localmarket, cultural, and regulatory knowledge, there-fore, is one of their priorities. Forming an alliancewith a local broker can thus achieve this goal.

Hypothesis 4c: Themoderating effect of a foreignfirm’s perceived capabilities will be stronger for theselection of local broker than for a centrally locatedlocal firm. In other words, foreign firms that havea high level of perceived capability are more likelyto select a local broker than a centrally located localfirm within the local alliance network.4

METHODOLOGY

Research SettingWe sampled the electronics and information tech-nology (IT) industries in China. We selected thesetwo industries for three reasons that are consistentwith our theoretical framework. First, open FDIpolicies in these two industries have attracted signi-ficant FDI (Sun, Chen, & Pleggenkuhle-Miles, 2010;Sun & Lee, 2013).5 Some local firms have developeda domestic alliance network, which allows them togain more bargaining power when negotiating withforeign firms for new IJVs (Meyer & Sinani, 2009).



Some local firms have honed the skill of conveyingthe value generated from their accumulation ofdomestic alliances to attract foreign firms. Second,the number of IJVs in these industries provides alarge sample. Third, some alliances may cross thesetwo industries because of the emerging trendof integrating hardware and software in globalcompetition.Meanwhile, we focused on alliance network rather

than other kinds of networks (e.g., board interlocks,value chain production networks, or patenting refer-ence groups). Methodologically, the type of contentof our choice of network should be linked with(1) resources that flow through the network and(2) the theory that we draw on (Wasserman & Faust,1994). The intuition would be that each type ofnetwork represents a specific set of pipes, throughwhich different resources flow. For example, valuechain production networks may be conduits forday-to-day operational flows, while alliances mayallow for strategic information and resource excha-nge. In our study, we are interested in the ability forforeign firms to obtain local market, cultural, andregulatory knowledge. The alliance network maybe a suitable venue to gain such business-relatedknowledge (Sun, Peng, Ren, & Yan, 2012). Whilepast research mainly focuses on the alliance at thedyad level (a focal firm’s allies make a difference), ourstudy centers on the alliance at the network level(the position occupied makes a difference).

Sample and Data CollectionFor the domestic alliance network, we focused onequity-based alliances, which require a serious com-mitment in resources. Alliance data were collectedfrom WIND Data Services. Similar to the SecurityData Corporation Platinum database, the WINDdatabase has reasonably consistent and completecoverage on alliance activities, and has been used inrecent research (Lin et al., 2009a; Shi, Sun, & Peng,2012; Sun & Lee, 2013).Following Rowley, Behrens, and Krackhardt

(2000), we constructed the domestic alliance net-work by incorporating all listed firms within twoindustries. In total, we identified 84 focal firms(listed on either the Shanghai or Shenzhen StockExchange) that had relatively complete financialinformation from WIND, resulting in a total of 191domestic alliances from 2001 to 2005 (inclusive).From WIND, we also obtained 73 IJV formationevents with 69 foreign partners among these 84local firms during these 5 years.6 Foreign partnersincluded Ericsson, GE, Hitachi, Hyundai, Intel, LG,

Microsoft, Motorola, NEC, Nokia, Philips, Samsung,Sony, Toshiba, and other world-class MNEs. Famouslocal partners included the world’s top TV makersTCL, Changhong, Konka, and Hi-Sense; the world’stop refrigerator maker Haier; the world’s top air-conditioner maker MIDEA; China’s largest TFT-LCDmaker BOE; IBM’s previous OEM Greatwall; China’stop software developer UFIDA; China’s largest soft-ware outsourcing providers Neusoft and ChinaSoft;China’s top PC maker Founder; and one of theworld’s top telecommunications equipment provi-ders ZTE. We also collected shareholder background,state ownership, and other corporate governanceand finance data from WIND.Local firms’ patent information was collected from

China’s State Intellectual Property Office.7 Localfirms’ CEO profiles were obtained from the “Profileof Directors and Senior Managers” section of theirannual reports. Foreign partners’ local market expe-rience data were collected from foreign partners’annual reports and news in Lexis-Nexis and theDow Jones News Retrieval Service. We obtained dataon country-level institutional indicators for foreignfirms from various editions of the World Competitive-ness Yearbook (2001–2005). We also accessed theFortune Global 500 Online Database to obtain dataon Fortune Global 500 membership data (http://money.cnn.com/magazines/fortune/global500/2008/).

Main Variables

The number of new foreign partnersOur interest was the likelihood of a local firm beingselected by a foreign firm as an IJV partner. Thelikelihood significantly increases when there are alarge number of new foreign firms that select localfirms as their local partners. Specifically, we countedthe number of new foreign partners that have chosena local Chinese firm as an IJV partner in a specific year.Alternatively, a simple way to operationalize our

dependent variable was to use a binary variable –

whether a domestic firm was selected as partner offoreign firms in IJVs (1) or not (0). While this did notcapture the number of new foreign partners, we hadsimilar results using a count and binary variable.8

Network centralityWe computed eigenvector centrality9 in the dome-stic alliance network using UCINET 6 (Borgatti,Everett, & Freeman, 2002), after constructing thesymmetric (non-directional) matrix for each year.A local firm obtains higher value of eigenvectorcentrality by being connected to a group of partners



http://money.cnn.com/magazines/fortune/global500/2008/

http://money.cnn.com/magazines/fortune/global500/2008/

that are themselves well connected (Jensen, 2008;Podolny, 2001; Shipilov & Li, 2008). Our formulafollows Bonacich (1972). Given an adjacency matrixA, the centrality of vertex i (denoted C1), is given by

Ci ¼ αX

AijCj (1)

where α is a parameter. The centrality of each vertexis therefore determined by the centrality of the verti-ces it is connected to. The parameter α is required togive the equations a non-trivial solution and istherefore the reciprocal of an eigenvalue. It followsthat the centralities will be the elements of thecorresponding eigenvector.

BrokerageBurt (1992) suggested that network constraint effec-tively measures a firm’s lack of access to structuralholes. Following this constraint formula, we have:

Pij +Xq

PiqPjq; q≠ i; j (2)

where Pij captures the strength of direct ties fromfirm i to firm j, and∑qPiqPjp is the sum of the indirectties’ strength from firm i to firm j via all firms q. If theconstraint score is non-zero, we calculate the broker-age as 1 minus the constraint score. If the constraintscore equals 0, it suggests the firm is not connectedto any others in the network (Jensen, 2008; Sun &Lee, 2013). In Figure 2, we illustrate the overallnetwork in 2005. We separate unconnected clustersfrom one another and move them to the right edgesof the figure.

Foreign firms’ local market experienceWe searched a foreign firm’s entry year from itsannual reports and news. Then we measured aforeign firm’s local market experience by the

difference between the selected year and the yearwhen the foreign firm entered the local market (Luo& Peng, 1999). We did not use a count-based experi-ence variable for foreign firms because our theoreti-cal arguments mainly focus on foreign firms’ localmarket experience (in the context of China) ratherthan general or specific alliance experience.

Foreign firms’ perceived capabilitiesSince foreign firms come from different countries, itis not easy to find common criteria to measure theircapabilities. Therefore local firms tend to look forcommon indicators of past performance. The For-tune Global 500 uses a common set of criteria acrossdifferent countries. The literature on certificationhas shown that certifications bestowed by thirdparties (e.g., Fortune Global 500 rankings) mitigatethe uncertainties faced by stakeholders when acces-sing a firm’s capabilities (Polidoro, 2013). This hasbeen empirically verified in a variety of settingssuch as education (Rindova, Williamson, Petkova, &Sever, 2005) and high-technology (Pollock & Gulati,2007). In an emerging economy such as Chinawhere information transmission is less transparent,local firms are more likely to embark on interna-tional standards as a common measure to evaluateforeign firms. In fact, there is enthusiastic reportingby the media on the Fortune Global 500 firms inChina. We, therefore, believed that these rankings(based on revenues and assets) offer a suitable proxyfor perceived capabilities of foreign firms (Lin, Yang,& Arya, 2009b). Thus we created a dummy vari-able for Fortune Global 500 membership (1) orotherwise (0).

Control Variables

Firm performanceGood past performance is likely to attract foreignpartners. Past performance thus was measured bythe average return on assets during the previous5 years.

Average domestic alliance sizeWe counted invested assets in all domestic alliancesannually and calculated their average value over thesize of the domestic alliance network. This informa-tion is available in CSRC filings and can be readilyobtained by foreign firms during the process ofdue diligence. Large average domestic alliance sizeenhances the information salience to the outsiders.Prior research shows that when information saliencereaches a threshold level, decision makers are moreFigure 2 Local alliance network in 2005.



likely to pay attention to and further integrate thisinformation into decision making (Davis & Ashton,2002).

Local firms’ international alliance experienceThis measure captures how effectively local firmscan manage their cross-border alliance (Glaister &Buckley, 1997). Following Luo and Peng (1999), werecorded the year a local firm established its firstinternational alliance, and then measured allianceexperience by the difference between the selected yearand the year when the first alliance was established. Inaddition, we also used count-based experience vari-ables by the number of previous international alliance(s) a local firm had prior to the time period of ourstudy (Gulati, Lavie, & Singh, 2009). Since we foundno significant difference, we only reported the count-based variable in the final result.

InnovationIn our two industries characterized by rapid inno-vation, a local firm’s capabilities of innovation areimportant (Khoury & Peng, 2011; Nielsen, 2003)and may affect the formation of IJVs. We measureda local firm’s innovation capabilities as the numberof patents filed in China each year (Ahuja, 2000).We used the logarithmic form to overcome dataoverdispersion.

Media exposureThe previous media exposure of alliances may affectpartner valuation and new IJV formation (Pollocket al., 2008). Using financial media for example,Pollock and Gulati (2007: 349) argued that “a firm’sability to obtain coverage may help to maintain thecompany’s visibility and cognitive availabilitybecause repeated exposure via analysts’ quarterlyearnings estimates, ratings and periodic reports andupdates makes the company more familiar and easierto recall.” Therefore we counted the number of searchresults when we searched for key words in Google(English) and Baidu (Chinese) search engines. Weused key words such as “alliance,” “cooperation,”“joint venture,” and “cooperative relationship”. TwoMBA students independently collected the news andcoded the exposure variable under a standardizedprotocol. The intercoder reliability was 95%. Wethen took the natural logarithm of the number ofmedium exposure.

Outside director ratioA high ratio of outside directors may affect afirm’s search for IJV partners and reduce inside

(executive) directors’ power in IJV formation (Gulati& Westphal 1999). Many outside directors at Chi-nese firms play this role (Peng, 2004). We capturedthis with the percentage of directors who are classi-fied as outside directors.

CEO dualityIn China, CEO duality has been found to be a double-edged sword, sometimes adding value but othertimes reducing the value of decision making (Peng,Li, Xie, & Su, 2010). Thus we created a dummyvariable if a CEO is chairman of board (1) or other-wise (0).

CEO overseas backgroundRecent studies suggest that local firm CEOs whohave international experience are familiar with Wes-tern-style managerial skills are more likely to bepreferred by foreign partners (Liu, Lu, Filatotchev,Buck, & Wright, 2010). We obtained CEO profilesfrom the annual reports that contained informationon their education, professional background, andcareer history. We created a dummy variable, CEOOverseas Experience, to reflect whether he or she hadworked or was currently working in a foreign MNE,an IJV, or an overseas branch office of a Chinesecompany (1), or not (0). We created another dummyvariable, CEO Overseas Education, to reflect whetherhe or she had been educated abroad (1) or not (0).

Institutional distanceInstitutional distance may affect a foreign firm’scosts and benefits of cooperating with local partners,and determine IJV formation (Xu & Shenkar, 2002).Following Gaur and Lu (2007), we used a Euclideandistance calculation on country-level indicatorsrelated to regulative and normative aspects of insti-tutional environment, to construct institutional dis-tances. This variable ranges between 0 and 2.3875.

State-owned enterprise (SOE)This dummy variable was measured by whethera firm’s controlling shareholder is an SOE or not.A controlling shareholder is one that holds a mini-mum of 25% of firm shares and has main influenceover firm decisions. Whether a local partner is anSOE may influence its attractiveness as a valuablepartner to foreign partners (Lu & Ma, 2008; Stan,Peng, & Bruton, 2013). Therefore we further inclu-ded State Ownership as a control variable. It wasmeasured by the percentage ownership owned bythe state.



Marketization indexAs a large emerging economy, China has substan-tial regional differences, thus making it necessaryto control for such a nuance (Kwon, 2012; Lu &Ma, 2008). Certain regions have developed bettermarket-supporting institutions, thus facilitatingthe dissemination of clear investment-related infor-mation (Peng, 2003). Following Fan, Wang, andZhu (2007a), we defined regions based on the pro-vinces in China. Fan et al. (2007a) developed thisprovince-based marketization index through fivedifferent categories: (1) government and marketforces; (2) development of non-SOEs; (3) develop-ment of commodity markets; (4) development offactor markets; and (5) development of market inter-mediaries and a legal environment. Each categoryis also further divided into several subcategories.This index has been used extensively in finance andeconomics (Chen, Firth, Gao, & Rui, 2006) as wellas management (Markóczy et al., 2013; Shi et al.,2012) and international business (Gao, Murray,Kotabe, & Lu, 2010). Since the score achieved byeach province changes over time, our marketizationindex is time-variant. Following Fan et al. (2007a),we included an index (Marketization Index) to capturethe stage of institutional progression in differentprovinces ranging from 2.95 to 10.41. A highermarketization index indicates a higher degree ofdevelopment of market-supporting institutions.

Political tiesThis variable was measured by determining whetherthe CEO of a domestic firm was or is an official of thecentral government, local government, or military.This measure is consistent with the one used inprior literature. For example, Fan, Wong, and Zhang(2007b) defined political ties as a CEO serving asa current or former government bureaucrat. As Fanet al. (2007b: 331) argued, because “Chinese govern-ment possesses the right to appoint the CEO of alisted company, the CEO’s political affiliation pro-vides a suitable proxy for government influence”. Asa result, ties with the government are often asso-ciated with greater resource support from the gov-ernment (Lu & Ma, 2008; Peng & Luo, 2000).Therefore firms with political ties may be perceivedas a good alliance candidate (Siegel, 2007).Other control variables included Firm Age (in

years), Firm Size (number of employees of the firm,using natural logarithm), Year Dummies, and IndustryDummies (which equals 0 if in the electronics indus-try and equals 1 if in IT industry).

Estimation StrategySince the dependent variable is a count variable (thenumber of foreign partners that the focal firm has),it ranges from 0 to a positive number, which is non-negative and makes it inappropriate to use standardmultiple regression. Poisson regression is explicitlydesigned for count dependent variables. However,Poisson assumes that the mean and variance of thecounts are equal. For most social-science data, thevariance likely exceeds the mean, resulting in theproblem of overdispersion that biases the estimatedstandard errors downward. The negative binomialmodel overcomes the overdispersion problem andalso accounts for omitted – we used a differentestimation and model that directly delivered thestandard error for the variable of interest variablebias. But it is only suitable when dependent variablesspan a wide range of values. In our study, thedependent variable – the number of foreign partners– ranges from 0 to 4. Under these conditions, ordinallogistic regression is more suitable. We, therefore,used a general linear model that incorporated ordi-nal logistic analysis. The model is:

Log P Y⩾a½ �=P Y<a½ �f g ¼ βa + β1X1

+ β2X2 + β3X3 + � � � + βpXp ð3Þ

We assumed that the response variable is ordinaland takes on integer values a=0, 1, 2, 3, 4 in oursample. If 1 ⩽Y⩽4 then we have four equations fora=0, 1, 2, 3. The Χ’s are covariates or interactions.This model is also called the proportional oddsmodel. In theory we could have four different coeffi-cients, β1,1, β1,2, β1,3, β1,4, one for each equation.But we have only one coefficient for Χ1. This isa simplified model that states that the odds forobtaining a response of at least a, for different valuesof a, is proportional to exp (βa),

10 whatever the fixedvalue of all the covariates. We used the GENMODprocedure in SAS.We had multiple observations for a firm over time,

which may raise the concern of potential autocorre-lation. To address this, we used a repeated measurein the GENMOD procedure that allowed us to con-trol for autocorrelation (AR1). This is akin to con-trolling for number of foreign partners in the prioryear. We also lagged independent variables andcontrol variables by 1 year in the regression analysis.To test Hypotheses 2b, 3c, and 4c, we needed

to compare the coefficient of network centralityand brokerage: βNetwork centrality vs βBrokerage. Becauseit was difficult to compute the standard error –

namely, the square root of the variance of the



differences between the two coefficients (e.g.,ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiβ̂Networkcentrality - β̂Brokerage

q) – we used a different esti-

mation and model that directly delivered the stan-dard error for the variable of interest (Wooldridge,2006). We first defined a new parameter θ so thatθ= βNetwork centrality−βBrokerage. Then, we tested a newnull hypothesis:

H0:θ ¼ 0 vsHbeing tested : θ<0 (4)

We can do this by rewriting the regression modelso that θ appears directly with one of the indepen-dent variables. Because βNetwork centrality= θ+ βBrokerage,we can plug this into our original model and rear-range the equation in the following way:

Log number of foreign partnersð Þ¼ βo + θ + βBrokerage

� �Network centrality

+ β Brokerage ´Brokerage + βcontrol variablesĆontrol variables + u ð5Þ

¼βo + θ Ńetwork centrality + βBrokerage´ Brokerage +Network centralityð Þ+ βcontrol variables Ćontrol variables + u ð6Þ

Equation (6) is essentially a different way of writ-ing Eq. (5). However, through estimating the coeffi-cient θ, we can identify the difference between thevalue of centrality and brokerage. This equation canalso be applied in comparison with the other twopairs of interactions.

RESULTSTable 1 presents descriptive statistics. Multicolli-nearity is not a significant problem – the averagevariation inflation factor, 1.57, is lower than thethreshold level of 10. Descriptive statistics showthat the IT industry has a 12% higher level ofalliance formation than the electronics industry.Both centrality and brokerage show positive correla-tion with the number of foreign partners, lendingpreliminary support to our Hypotheses 1 and 2a.Table 2 shows the results of the ordinal logistic

regression. Model 1 is the base model. Hypothesis 1asuggests that network centrality is positively rela-ted to the likelihood of being selected by foreignfirms. Model 2 supports Hypothesis 1a (p<0.001).This means that controlling for all other factors, theodds for Y (being selected by N number of foreignfirms, for N=1, 2, 3, 4, per year) are predicted by ourmodel to increase by a factor of 3.062 when network

centrality is increased by one unit.11 Hypothesis 1bis not supported.Hypotheses 3a, 3b, 4a, and 4b are also supported.

Hypotheses 3a and 3b show stronger significancethan Hypotheses 4a and 4b. These hypotheses arguethat the relationship between network positiontraits and the likelihood of local firms being selectedby foreign firms will differ by (1) whether foreignfirms have sufficient local market experience and(2) whether these foreign firms are perceived by localfirms as resource-abundant. These hypotheses aresupported in Models 3, 4, 5, and 6, respectively.Next, following Aiken and West (1991), we explo-

red the significant two-way interactions as reportedin Table 3. For each significant interaction wecalculated the simple slopes of network centrality/brokerage on the likelihood of being selected byforeign firms and their standard errors at three levels(i.e., mean, one standard deviation above, and onestandard deviation below the mean) of the secondpredictors (i.e., average size of alliance, foreign part-ner’s local market experience, and foreign partner’sFortune 500 membership) as suggested by Cohenand Cohen (1983). We then conducted t-tests on thevalues of the simple slopes divided by their standarderrors. Table 3 indicates that the positive relation-ship between network centrality and the likelihoodof being selected by foreign firms is stronger whenforeign firms have extensive local market experiencethan when they have limited experience.Table 4 reports the comparison test of coeffi-

cients. In Model 7, we compared the coefficientsof network centrality and brokerage. Positive andsignificant coefficients on network centrality inModel 7 indicate the coefficient of centrality is sig-nificantly larger than that of brokerage, providingsupport for Hypothesis 2. Models 8 and 9 havenegative and significant coefficients on centrality,suggesting that the coefficient of network centralityis significantly smaller than that of brokerage inboth interaction terms. These results lend supportfor Hypotheses 3c and 4c, respectively.

Robustness ChecksOur estimates would be unbiased only if self-selec-tion bias and omitted variable bias were accountedfor. In other words, local firms’ decisions to ally needto be random and we need to account for unobser-ved factors that might impact these decisions.To assess these endogeneity issues, we employeda two-stage Heckman (1979) selection procedure.The results were consistent with the reported find-ings and the inverse Mills ratio was not significant,



Table 1 Descriptive statistics and Pearson correlation matrix (N=323)

Variables Mean Standarddeviation

Minimum Maximum 1 2 3 4 5 6 7 8 9 10 11

1 Number of foreign partners 0.19 0.59 0 4 12 Firm age 9.08 4.69 1 22 0.02 13 Firm size 7.23 1.20 4.09 11.09 0.05 −0.04 14 Performance 0.26 11.65 −135.38 27.93 0.09 −0.24* 0.13* 15 Patents 0.30 0.53 0 2.11 −0.04 −0.14* 0.34* 0.10 16 Media 3.98 5.17 0 13.79 0.07 0.04 0.23* −0.04 0.14* 17 Industry 0.67 0.47 0 1 0.12* 0.14* −0.37* −0.06 −0.15* −0.18* 18 Outside director ratio 0.26 0.13 0 0.5 −0.23* 0.16* −0.02 −0.18* 0.21* 0.05 −0.01 19 CEO duality 0.16 0.37 0 1 −0.02 0.06 0.04 −0.09 −0.03 0.08 −0.09 −0.04 1

10 CEO oversea experience 0.15 0.36 0 1 −0.04 −0.10 −0.09 0.05 0.02 0.04 0.01 −0.04 −0.17* 111 CEO oversea education 0.05 0.21 0 1 −0.05 −0.16* −0.04 0.01 0.00 −0.04 −0.05 −0.08 −0.09 0.29* 112 Institutional distance 0.19 0.55 0 2.39 0.10* −0.07 −0.11* 0.05 −0.04 0.01 −0.03 0.01 −0.02 −0.09 −0.0413 Average domestic alliance size 4947 8830 0 99904 −0.04 0.02 0.04 −0.04 −0.03 0.01 −0.17* −0.01 −0.10 −0.06 −0.0214 Local firms’ foreign experience 2.01 2.37 0 7 0.07 −0.08 0.22* 0.02 0.16* 0.24* −0.35* 0.11* −0.05 −0.14* −0.0915 Foreign firms’ local market

experience1.72 5.89 0 39 0.03 −0.08 −0.01 −0.04 0.02 0.41* −0.12* 0.00 −0.05 0.06 0.04

16 Fortune 500 membership 0.05 0.22 0 1 0.07 −0.08 −0.02 −0.05 0.05 0.31* 0.04 0.05 −0.10 0.15* 0.0717 SOE 0.73 0.44 0 1 0.09 −0.06 0.23* 0.21* 0.11* 0.11* −0.24* −0.15* 0.01 0.10 0.0118 Marketization index 7.54 1.71 2.95 10.41 −0.02 0.21* 0.20* −0.02 0.25* 0.29* 0.07 0.35* −0.05 0.10 −0.0119 Political ties 0.17 0.38 0.00 1.00 0.00 −0.19* −0.04 0.03 −0.01 0.00 0.04 −0.02 0.29* −0.03 −0.1020 State ownership 30.02 26.43 0.00 85.00 0.08 −0.38* 0.13* 0.14* 0.07 0.04 −0.13* −0.13* −0.06 0.18* 0.0421 Network centrality 0.36 0.82 0 6.3 0.35* −0.04 0.10* 0.03 −0.03 0.05 −0.13* −0.18* −0.04 −0.04 −0.0422 Brokerage 0.28 0.45 0 1 0.28* 0.05 0.02 −0.01 0.03 0.02 0.15* −0.16* −0.05 −0.09 −0.05

12 13 14 15 16 17 18 19 20 21 22

12 Institutional distance 113 Average domestic alliance size −0.04 114 Local firms’ foreign experience −0.03 0.02 115 Foreign firms’ local market experience −0.04 0.21* 0.38* 116 Fortune 500 membership −0.02 0.09 0.08 0.33* 117 SOE −0.05 0.04 0.27* 0.04 −0.03 118 marketization index −0.05 −0.07 0.09 0.17* 0.17* −0.11* 119 Political ties 0.00 −0.10 −0.04 −0.10* −0.04 −0.04 0.05 120 State ownership −0.04 0.17* 0.26* 0.13* 0.01 0.59* −0.27* −0.08 121 Network centrality 0.00 0.02 0.01 −0.01 −0.02 0.03 −0.04 −0.04 0.03 122 Brokerage 0.03 −0.01 −0.03 −0.02 −0.03 0.01 0.00 0.01 −0.06 0.38* 1

Note: * p<05.

Domestic

alliance

netw

ork

toattract

foreig

npartn

ersWeilei(Stone)

Shietal

352

JournalofInternationalBusinessStudies

Table 2 Ordinal logistic regression analysis by general linear models

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Control VariablesFirm age 0.018 0.021 0.033 0.038 0.027 0.028Firm size 0.392** 0.324* 0.346* 0.337* 0.348* 0.334*Performance 0.002 0.002 0.001 0.001 0.001 0.001Patents −0.109 0.131 −0.023 0.132 −0.013 0.095Media 0.047 0.033 0.031 0.033 0.034 0.032Industry 1.303*** 2.321*** 2.235*** 2.328*** 2.345*** 2.338***Outside director ratio −4.617* −2.212 −3.336 −3.631 −2.646 −3.159CEO duality 0.074 0.736† 0.813† 0.736 0.644 0.636CEO oversea experience −0.459 −0.002 −0.223 −0.063 −0.213 −0.089CEO oversea education −1.114 −0.315 0.034 0.163 −0.146 −0.141Institutional distance (ID) 0.567† 0.737* 0.759* 0.768* 0.756* 0.771*Average domestic alliance size (ADAZ) 0.116 0.093 0.071 0.086 0.082 0.091Local firms’ foreign experience (LFE) 0.185* 0.325** 0.348** 0.224** 0.227** 0.238**Foreign firms’ local market experience (FLE) −0.048 −0.037 −0.128 −0.167* −0.075 −0.082Fortune 500 membership (F500) 1.897* 2.238† 4.017** 3.786* 2.563 2.517SOE 0.133 0.113 0.014 0.156 0.146 0.154Marketization index −0.074 −0.064 −0.135 −0.056 −0.061 −0.067CEO political ties 2.231* 2.735* 2.926* 2.526† 2.837* 2.463†

State ownership 0.013 0.014 0.026† 0.033† 0.031† 0.032†

Main effectNetwork centrality (Hypothesis 1a+) 1.119*** 1.298*** 1.154*** 1.171*** 1.138***Brokerage (Hypothesis 1b+) 0.507 0.261 0.362 0.284 0.327

Interaction effectNetwork centrality×FLE (Hypothesis 3a+) 0.391**Brokerage×FLE (Hypothesis 3b+) 0.776**Brokerage×F500 (Hypothesis 4a+) 6.549*Network centrality×F500 (Hypothesis 4b+) 12.617†

Intercept 1 −9.425*** −10.549*** −10.926*** −10.853*** −10.942*** −10.636***Intercept 2 −8.992*** −10.132*** −10.365*** −10.345*** −10.458*** −10.153***Intercept 3 −7.873*** −9.236*** −9.766*** −9.467*** −9.438*** −9.224***Intercept 4 −6.675*** −7.254*** −7.378*** −7.435*** −7.576*** −7.306***N 323 323 323 323 323 323Log likelihood −166.30 −149.25 −137.38 −137.11 −138.28 −138.34

Notes: †p<0.10, *p<0.05, **p<0.01, ***p<0.001.Due to the space limitation, we omit the coefficient estimates of the year dummy variables in this table.

Table 3 Results of standard error and t−tests for simple slopes of two-way interactions including network position traits and secondpredictors

Network centrality Brokerage

Simple slope Standard error t-test Intercept Simple slope Standard error t-test Intercept

Foreign firm’s local market experienceHigh 1.863 0.382 4.88*** −12.801** 1.285 0.448 2.87** −12.614***Mean 1.298 0.291 4.46*** −11.338*** 0.362 0.511 0.71 −10.937***Low 0.702 0.301 2.33* −10.425*** −0.667 0.685 −0.97 −9.919***

Fortune 500 membershipHigh (member) 1.434 0.347 4.13*** −11.453*** 0.801 0.463 1.73† −10.962***Mean 1.171 0.286 4.09*** −11.275*** 0.327 0.556 0.59 −10.857***Low (not a member) 0.764 0.283 2.70** −12.436*** −0.103 0.686 −0.15 −12.016***

Note: †p<0.10, *p<0.05, **p<0.01, ***p<0.001.



indicating that endogeneity did not significantlyimpact our results. In addition, we also tested a fixedeffect model, which generated similar results.

DISCUSSION

ContributionsIn our view, at least two contributions emerge. First,our theoretical framework enriches research oninterorganizational relationships, especially on IJVformation. We find that certain network positiontraits within the alliance network seem to be asso-ciated with a greater likelihood of participating in anIJV. Theoretically, network position traits as selec-tion criteria represent an under-researched area inthe partner selection literature (Luo, 1998). We

investigate the importance of network position traitsin attracting foreign IJV partners. As foreign investorsbecome strategic insiders (Luo, 2007), they not onlyseek a local partner’s capability attributes (opera-tional, technological, and organizational), but alsomore critically pay attention to a local partner’sdomestic network. While prior research offers exten-sive knowledge of the role of networks in partnerselection, recent research has begun to explore therole of centrality and brokerage simultaneously infirms’ cross-border investment decisions (Guler &Guillén, 2010). Extending this line of research, weargue that brokerage can be an important attributethat affects foreign firms’ market entry. The value ofbrokerage is also particularly important in emergingeconomies where regulatory regimes are less well

Table 4 Ordinal logistic regression analysis by general linear models on comparison

Model 7 Model 8 Model 9

Control variablesFirm age 0.023 0.035 0.027Firm size 0.314* 0.363* 0.339*Performance 0.002 0.001 0.001Patents 0.131 −0.018 −0.011Media 0.025 0.027 0.031Industry 2.221*** 2.294*** 2.369***Outside director ratio −2.363 −3.629 −2.667CEO duality 0.784† 0.862† 0.656CEO oversea experience −0.002 −0.218 −0.232CEO oversea education −0.515 0.062 −0.173Institutional distance (ID) 0.727* 0.801* 0.797*Average domestic alliance size (ADAZ) 0.093 0.071 0.083Local firms’ foreign experience (LFE) 0.257** 0.242** 0.233**Foreign firms’ local market experience (FLE) −0.063 −0.152 −0.068Fortune 500 membership (F500) 2.153† 4.147** 2.573SOE 0.152 0.028 0.135Market index −0.049 −0.095 −0.078CEO political ties 2.863* 2.945* 2.924*State ownership 0.012 0.036† 0.035†

Main effectNetwork centrality (Hypothesis 2+) 0.629* 1.418* 1.308**Brokerage+network centrality 0.259 −1.403 −1.418

Interaction EffectNetwork centrality×FLE (Hypothesis 3c−) −1.064*(Network centrality+brokerage)×FLE 0.938*Network centrality×F500 (Hypothesis 4c−) −38.478***(Brokerage+Network centrality)×F500 30.002***

Intercept 1 −10.637*** −10.939*** −10.812***Intercept 2 −10.319*** −10.439*** −10.343***Intercept 3 −9.263*** −9.567*** −9.426***Intercept 4 −7.383*** −7.649*** −7.419***N 323 323 323Log likelihood −149.64 −137.96 −137.61

Note: †p<0.10, *p<0.05, **p<0.01, ***p<0.001.



established, where uncertainties are high, yet wheresignificant risks are accompanied by large marketrewards (Li et al., 2009; Li et al., 2013). In such anenvironment, competitive advantages may accruefor foreign firms that can access tacit local “recipes”since such local information and knowledge arenot readily available on the open market. WhileGuler and Guillén (2010) focused on centrality fromthe perspective of foreign partners, we go aboveand beyond their work by extending our attentionto cover brokerage advantage from the perspective ofdomestic partners.Second, the comparative analysis between cen-

trally located local firms and local brokers has helpedus to develop the boundary conditions of networktheory – when and under what conditions networkposition traits are more likely to be observed andinterpreted. For example, Pollock and Gulati (2007)pointed out that existing studies cannot differentiatethe relative benefits of each network position trait interms of visibility enhancement and uncertaintyreduction. Our results indicate that centrally locatedlocal firms and local brokers clearly convey differentlevels of visibility. This is particularly important ina highly dynamic and uncertain environment(Lavie, 2007) such as in emerging economies, wherethe importance of alliances for local knowledgeacquisition and legitimacy establishment are bothespecially significant. In such an environment, net-work position traits that may have the visibility-enhancing and uncertainty-reducing capabilitiesmay well influence firms’ partnering behaviors indifferent ways (Pollock et al., 2008).More importantly, our comparative analysis of

centrality and brokerage contributes to the alliancenetwork literature. Prior studies have found thatfirms’ centrality in their alliance network is animportant network position trait that attracts poten-tial partners’ attention (Benjamin & Podolny, 1999).Some studies have focused on how brokerage canconstrain a local broker’s ability to be selected asa partner (Jensen, 2008). Few studies have comparedthe relative importance of each network positiontrait. We argue that foreign firms’ local marketexperience and their capabilities as perceived by thelocal firms constitute important boundary condi-tions that suggest one type of network position traitmay be more important than the other. Guler andGuillén (2010) found that firms that have highcentrality in the home country network tend toenjoy a higher rate of foreign market entry. One ofthe key assumptions underlying this finding is thatforeign firms differ in terms of their home-country

network characteristics. This heterogeneity assump-tion makes local partners’ network position traitparticularly important for foreign partners. Accor-ding to Guler and Guillén (2010), foreign firmsthat already outperform other firms in their homecountries in social status would be willing to act ondissimilar capabilities of local partners firms. In fact,this type of “matching” argument is consistent withour Hypothesis 4c, where we argue that foreign firmsthat have a high level of perceived capabilities aremore likely to select a local broker than a centrallylocated local firm. In our study, foreign firms’ per-ceived capabilities are measured as a function of theforeign firms being a Fortune Global 500 member ornot. Conceptually, social status is the prestigeaccorded actors due to their social positions withinthe system. Empirically, a suitable measure would bethose related to a network such as centrality (Guler &Guillén, 2010). Jensen and Roy (2008) used analternative measure, whether a firm belongs to theBig 4 Accounting Firm to operationalize social status.To this end, the Fortune Global 500 membershipcan be viewed as an alternative measure for socialstatus, although we agree that this is not a perfectmeasure. In a nutshell, local brokers usually possesscomplementary (but non-overlapping) knowledgesuch as unique local market knowledge (Shipilov &Li, 2008). Therefore Fortune Global 500 firms mayfind a local broker more attractive as an IJV partnerrelative to a centrally located local firm.

Managerial and Policy ImplicationsOur findings have two important implications formanagers and policymakers. Our results indicatethat managers of foreign firms should use cautionwhen applying a general theory of selecting an IJVpartner based on the local firm’s location in thedomestic network. Instead, foreign firms’ managersmust be aware of the context of their decision, andconsider the contingent implications of centrallylocated local firms vis-à-vis local brokers in the dome-stic market. For our study, centrally located localfirms and local brokers offer different strategicresources. The longer the foreign firm is in themarket, the more likely it is to select a local brokeras an IJV partner. Anticipating this search processand acquiring the “best” local broker IJV partnermay prove critical for the long-term success in thedomestic market.Policymakers may also find value in knowing how

foreign firms select local partners. Often, policy-makers in emerging economies will replicate insti-tutions found in developed economies. This is



problematic because the resulting institutions maymislead foreign firms looking for information onlocal partners (Spicer, Dunfee, & Bailey, 2004).Instead, policymakers may look for ways to enhancethe quality of network position traits from localfirms in domestically unique ways – in this casereporting whether the firm is a centrally locatedlocal firm or a local broker. By creating a publicforum for local firms to display their network posi-tion trait, policymakers may create social pressuresto avoid misrepresentation. This will increase thecosts associated with imitation, making the net-work position trait easier to identify and interpretfrom the perspective of the foreign firm. Overall,such a system may facilitate the partner selectionprocess and (potentially) attract more FDI to thecountry by mitigating information asymmetries(Stiglitz, 2000).

Limitations and Future Research DirectionsOne of the limitations is that we do not measurewhether foreign partners truly map out local com-pany networks and the position of specific compa-nies within them when making partnering deci-sions. To address this issue, we did three things tofurther fine-tune our arguments. First, Godfrey andHill (1995) argued that simply because we cannotobserve certain phenomena does not mean theydo not exist. “The evidence that we have that suchentities exist independent of our theorizing aboutthem is not based upon observation of the entitiesthemselves, since they are unobservable, but uponobservation of their effects” (Godfrey & Hill, 1995:525). In other words, we believe the identificationprocess is essentially an underlying mechanism,which is unobservable. However, we can documentits existence by empirically observing its effects onpartner selection.Second, theoretically, individuals are highly moti-

vated to generate an overall view of the system (“theforest”) when they enter into a new social situation(von Hecker, 1993). They are eager to learn andto make sense of the complex network structures orpatterns. Within the alliance literature, some practi-tioners pointed out the importance of understand-ing “group structure” of alliance partners. As Wolf(2000: 340) stated, “We therefore must under-stand all the major commercial relations … that thejoint venture company has with other enterprises.This requires an inquiry into the interrelation-ships between the joint venture company andothers. This inquiry will afford a comprehensiveunderstanding of the dynamics of the joint venture

company.” Furthermore, the fact that high socialstatus foreign firms will be more likely to search forlocal brokers rather than centrally located local firms(Hypothesis 4c) further reinforces the notion thatsome (although not all) foreign firms are indeedobserving these network position traits.Third, to further enhance the face validity of MNEs

mapping local firms’ network, we conducted threeinformal interviews. As our interviewees pointedout, they do not map out the local firms’ network inthe same way that researchers do (i.e., calculatingnetwork centrality or brokerage). It is also impossibleor very challenging for boundedly rational people tokeep track of all possible relationship pairs (Kilduff,Crossland, Tsai, & Krackhardt, 2008). However, theinformation on who are central players and who arelocal brokers can be discerned through observing thebehavior clue or pattern of local firms and are sharedand agreed upon by local communities. For example,as one of our interviewees responded when asked“Do you truly pay attention to your potential alli-ance partners’ network?”

That (their network) is certainly important to us… you knowthe behavior patterns are important to decide what are thefirms that enjoy high social status here and what are thosebrokers …

Another interviewee pointed out:

Central players are usually easy to spot. Our suppliers talkabout them, so do our distributors and our legal and financialconsultants.

Another limitation is that because we examine IJVpartner selection in a single country, we cannotstudy the effect of different institutional effectson partner selection as Hitt et al. (2004) and Meyeret al. (2009a) did by comparing multiple countries. Apossible future direction is to compare how networkposition traits impact partner selection in multipleemerging economies with different market-support-ing institutions (Meyer et al., 2009a). Similarly, ourexamination period is between 2001 and 2005, thefirst period after China’s accession to the WorldTrade Organization that resulted in an influx of FDI.To this end, our study hinges on an importantassumption: having foreign IJV partners is valuablefor local firms. However, the relative value of estab-lishing IJVs may decline over time (Pajunen & Fang,2013; Tong et al., 2008), especially after the estab-lishment of market-supporting institutions (Peng,2003; Xia, Tan, & Tan, 2008). Unfortunately, ourdata preclude us from investigating such a dynamiceffect, which should be explored in future research.



One of the limitations is that our political ties onlymeasure whether the CEO of a domestic firm was oris an official of the central government, local govern-ment, or military, but it does not capture the specificcommunist party links that are sometimes moreintricate and complex. For instance, the leadingcadres of the central government may follow theimplicit rule of intergenerational inheritance whenappointing top managers (Shi et al., 2013). Firmsthat are represented by these managers may be wellsought after by foreign partners due to these man-agers’ particularistic relations (Luo & Chung, 2005)with the top leaders in the government sector. Morenuanced data showing these differences can shedmore light on the role of political ties in facilitatingand constraining foreign partners’ selection.Also, our sample is limited to IJVs. We do not

include other types of alliances, especially non-equityalliances. This may limit the generalizability of ourarguments in other contexts. The alliance networksformed by equity-based IJVs may be different fromthose formed by non-equity-based alliances. Partnersin equity-based IJVs entail significantly more resourcecommitments. Therefore an alliance network formedby equity-based IJVs can be considered a networkwith multiple strong ties. Non-equity-based alliancesaremore likely to be characterized byweak ties, whichmay be even harder for foreign partners to identify.Future research may examine how the strength ofties impacts the observation and interpretation ofnetwork position traits.Furthermore, we only include listed firms. Our

main rationales for this sampling strategy are two-fold. First, publicly listed firms have much bettercorporate governance than private firms in China.Listed firms with sophisticated governance struc-tures may more strongly protect foreign firms’ inter-est than private firms. Second, publicly listed firmshave relatively more transparent policies than pri-vate firms. The more available corporate informationrelated to publicly listed firms will help foreign firms’partner selection. While private firms are active informing domestic alliance networks and some haveattracted foreign IJV partners, lack of data availabi-lity for private firms may pose a potential threatto the validity of our results. We, therefore, haveconducted some additional tests. In an ad hoc ana-lysis, we compare centrality and brokerage measuresbetween some listed firms and some private firms.The rationale is that if both groups demonstratedifferent levels of centrality and brokerage, thenthere is a potential sample selection bias. Our testindicates that there is no significant difference at the

level of centrality and brokerage between these twogroups.

CONCLUSIONPrior research on IJV partner selection has leftunexplored the potentially important network pro-perties of the domestic alliance network in whichlocal firms are embedded. We argue and substantiatethe case that the domestic alliance network can bea credible and powerful source for foreign firms tofilter and sort prospective local partners in an emer-ging economy. For practitioners, the implicationsof our findings are clear. Specifically, local firms’network position traits in the domestic alliance net-work can help foreign firms reduce informationasymmetries when selecting IJV partners. In turn,managers of local firms are advised to leverage theirdomestic alliance network to attract foreign IJVpartners by considering not only the foreign firms’capabilities, but also the foreign firms’ local marketexperience.

ACKNOWLEDGEMENTSWe thank Yadong Luo and three reviewers for editorialguidance, John Prescott and Ravindranath Madhavanfor helpful comments, and Lily Dong for discussion ofindustry insights. We also appreciate helpful commentsfrom the seminar participants at Weissman Center forInternational Business at City University of New York,the management departments at Zicklin School ofBusiness, Baruch College, City University of New York,and China Europe International Business School. Earlierversion of this manuscript was presented at the Acad-emy of Management (Boston, August 2012). Thisresearch was supported by the PSC-CUNY Award(63173-00-41), jointly funded by the Professional StaffCongress and the City University of New York, theWeissman Center International Business ResearchGrant, Kemper Summer Research Grant, funded byBloch School of Management, University of Missouri,Kansas City, the Jindal Chair at University of Texas,Dallas, and the National Natural Science Foundation ofChina (71132006). All views are those of the authorsand not necessarily those of the NSFC.

NOTES1While China has been plagued by information

credibility historically, the issue has been generallyimproving over time. For specific information aboutdisclosure requirements for listed companies, check thewebsite of China Securities Regulatory Commission(http://www.csrc.gov.cn/pub/csrc_en/).



http://www.csrc.gov.cn/pub/csrc_en/

2Foreign firms can de-code local media throughtwo main mechanisms. First, for the human mecha-nism, foreign firms can hire someone who understandsthe local language to de-code media. Many largeforeign firms also have a large ethnic group ofemployees (e.g., overseas Chinese), who can be usefulfor understanding local media. Second, for the socialmechanism, many local medium (e.g., newspapers,trade journals, TV programmers) now have an Englishversion (a global language) of their local content.This is particularly true in China where the forceof globalization makes such an endeavor necessary toattract FDI.

3For example, Japan’s Hitachi, which entered Chinaearlier, had more local market experience than theUnited States’ Corning did. Hitachi chose the FujianElectric Group in Fuzhou as the partner to build a digitalmedia JV company in 2001, and Corning selected SEGas the partner to build a JV for manufacturing colorcathode ray tubes and monitor bulbs. Fujian ElectricGroup was a local broker, while SEG was a centrallylocated firm in the local alliance network.

4The central element here would be that theforeign firm’s choice (local broker vs centrally locatedfirm) is not exclusively determined by the foreignfirm’s capability level. The capability level is one factorthat our theory proposes adds explanatory power –

we do not claim that it will exhaustively predict theforeign firm’s choice. Rather, what capability leveldoes is to predispose the foreign firm in some ways,not force that choice. In other words, the model isnot a “causal-deterministic” one but an “associative-likelihood” one.

5For three reasons, we focused on the FDI policyduring the study period (2001–2005). First, China’sPeople’s Parliament Committee revised the ForeignJoint Venture Law and Foreign Direct InvestmentLaw in 2000 to be more consistent with the WTOagreement. Second, the new FDI policy eliminatesthose restrictions that are incompatible with the WTOagreement. Third, the Chinese government also revisedindustry policies in order to attract FDI toward strategicindustries such as electronics and IT.

6One IJV event may engage two or more partners,since some foreign entrants (e.g., Microsoft) have twoor more IJVs with multiple local partners in China.

7We accessed http://search.sipo.gov.cn/sipo/zljs/. Insummary, China’s patent law divides patents into threecategories: invention, utility model, and externaldesign. Invention patents are regarded as majorinnovations. To obtain a patent for invention, anapplication must meet the requirements of “novelty,inventiveness, and practical applicability.” Usually, ittakes about 1–1.5 years for the State Patent Office toprocess an invention patent application. Theprocessing time for patent applications for a utilitymodel is about 6 months, and it is even shorter fordesign patent. Today’s patent law in China is mostly inline with the international standard. Up till now, Chinahas acceded to all the international patent treaties inorder to meet the requirements of the WTO’sAgreement on Trade-Related Intellectual Properties(Cheung & Lin, 2004).

8This approach generates similar results. Results areavailable upon request.

9We thank one reviewer for pointing this out.10We also conducted the Proportional Odds

Assumption test, which tests whether our one-equation model is valid (the ordered logit modelestimates one equation over all levels of the dependentvariable). If we were to reject the null hypothesis, wewould conclude that ordered logit coefficients are notequal across the levels of the outcome and we would fita less restrictive model. If we fail to reject the nullhypothesis, we conclude that the assumption holds.We first download a user-written command in Statacalled omodel. We then conduct likelihood ratio test(Long & Freese, 2006). The results (Prob> χ2=0.3427)indicate that we cannot reject the null hypothesis,which suggests that there is no difference in thecoefficients among models.

11We take the following form to calculate the Oddsratio.

Odds ¼Odds Y ⩾ a½ �whenXcentrality ¼ i +1Odds Y ⩾ a½ �whenXcentrality ¼ i

¼ expð1:119Þ�3:062

REFERENCESAhuja, G. 2000. Collaboration networks, structural holes,and innovation: A longitudinal study. Administrative ScienceQuarterly, 45(3): 425–455.

Ahuja, G., Polidoro, F., & Mitchell, W. 2009. Structural homophilyor social asymmetry? The formation of alliances by poorlyembedded firms. Strategic Management Journal, 30(9): 941–958.

Aiken, L. S., & West, S. G. 1991. Multiple regression: Testing andinterpreting interactions. Newbury Park, CA: Sage.

Al-Laham, A., & Souitaris, V. 2008. Network embeddednessand new-venture internationalization: Analyzing internationallinkages in the German biotech industry. Journal of BusinessVenturing, 23(5): 567–586.

Beckman, C. M., Haunschild, P. R., & Phillips, D. J. 2004. Friendsor strangers? Firm-specific uncertainty, market uncertainty,and network partner selection. Organization Science, 15(3):259–275.



http://search.sipo.gov.cn/sipo/zljs/

Benjamin, B. A., & Podolny, J. M. 1999. Status, quality, and socialorder in the California wine industry. Administrative ScienceQuarterly, 44(3): 563–589.

Bjerregaard, T., & Lauring, J. 2012. Entrepreneurship as institu-tional change: Strategies of bridging institutional contradic-tions. European Management Review, 9(1): 31–43.

Bonacich, P. 1972. Factoring and weighting approaches to statusscores and clique identification. Journal of Mathematical Sociol-ogy, 2(1): 113–120.

Borgatti, S. P., Everett, M. G., & Freeman, L. C. 2002. Ucinet 6 forwindows: Software for social network analysis. Harvard, MA:Analytic Technologies.

Brouthers, L. E., & Xu, K. 2002. Product stereotypes, strategy andperformance satisfaction: The case of Chinese exporters. Journalof International Business Studies, 33(4): 657–677.

Burt, R. S. 1992. Structural holes: The social structure of competition.Cambridge, MA: Harvard University Press.

Burt, R. S. 2005. Brokerage and closure. New York: OxfordUniversity Press.

Burt, R. S. 2007. Secondhand brokerage: Evidence on theimportance of local structure for managers, bankers, andanalysts. Academy of Management Journal, 50(1): 119–148.

Carter, A. P. 1989. Knowhow trading as economic exchange.Research Policy, 18(3): 155–163.

Chabowski, B. R., Hult, G. T. M., Kiyak, T., &Mena, J. A. 2010. Thestructure of JIBS’s social network and the relevance of intra-country variation: A typology for future research. Journal ofInternational Business Studies, 41(5): 925–934.

Chang, S. J., & Xu, D. 2008. Spillovers and competition amongforeign and local firms in China. Strategic Management Journal,29(5): 495–518.

Chang, Y.-Y., Gong, Y., & Peng, M. W. 2012. Expatriate knowledgetransfer, subsidiary absorptive capacity, and subsidiary perfor-mance. Academy of Management Journal, 55(4): 927–948.

Chen, G., Firth, M., Gao, D. N., & Rui, O. M. 2006. Ownershipstructure, corporate governance, and fraud: Evidence fromChina. Journal of Corporate Finance, 12(3): 424–448.

Cheung, K.-Y., & Lin, P. 2004. Spillover effects of FDI on innova-tion in China: Evidence from the provincial data. China Eco-nomic Review, 15(1): 25–44.

Choi, C.-B., & Beamish, P. W. 2013. Resource complementarityand international joint venture performance in Korea. AsiaPacific Journal of Management, 30(2): 561–576.

Cohen, J., & Cohen, P. 1983. Applied multiple regression/correlationanalysis for the behavioral sciences. Hallstale, NJ: LawrenceErlbaum.

Dacin, M. T., Oliver, C., & Roy, J.-P. 2007. The legitimacy ofstrategic alliances: An institutional perspective. Strategic Mana-gement Journal, 28(2): 169–187.

Damanpour, F., Devece, C., Chen, C. C., & Pothukuchi, V. 2012.Organizational culture and partner interaction in the manage-ment of international joint ventures in India. Asia Pacific Journalof Management, 28(2): 169–187.

Davis, E. B., & Ashton, R. H. 2002. Threshold adjustment inresponse to asymmetric loss functions: The case of auditors’‘substantial doubt’ thresholds. Organizational Behavior andHuman Decision Processes, 89(2): 1082–1099.

Dhanaraj, C., Lyles, M. A., Steensma, H. K., & Tihanyi, L. 2004.Managing tacit and explicit knowledge transfer in IJVs: The roleof relational embeddedness and the impact on performance.Journal of International Business Studies, 35(5): 428–442.

Ding, X.-H., Huang, R.-H., & Liu, D.-L. 2012. Resource allocationfor open and hidden learning in learning alliances. Asia PacificJournal of Management, 29(1): 103–127.

Fan, G., Wang, X. L., & Zhu, H. P. 2007a. Marketization index inChina: The regional process report of 2006. Beijing: EconomicScience Press, (in Chinese).

Fan, J. P. H., Wong, T. J., & Zhang, T. 2007b. Politically connectedCEOs, corporate governance, and post-IPO performance ofChina’s newly partially privatized firms. Journal of FinancialEconomics, 84(2): 330–357.

Figueiredo, P. N., & Brito, K. 2011. The innovation performanceof MNE subsidiaries and local embeddedness: Evidence froman emerging economy. Journal of Evolutionary Economics, 21(1):141–165.

Galunic, C., Ertug, G., & Gargiulo, M. 2012. The positiveexternalities of social capital: Benefiting from senior brokers.Academy of Management Journal, 55(5): 1213–1231.

Gao, G. Y., Murray, J. Y., Kotabe, M., & Lu, J. 2010. A “strategytripod” perspective on export behaviors: Evidence from domes-tic and foreign firms based in an emerging economy. Journal ofInternational Business Studies, 21(3): 141–165.

Gaur, A. S., & Lu, J. W. 2007. Ownership strategies and survival offoreign subsidiaries: Impacts of institutional distance andexperience. Journal of Management, 33(1): 84–110.

Geringer, J. M. 1991. Strategic determinants of partner selectioncriteria in international joint ventures. Journal of InternationalBusiness Studies, 22(1): 41–62.

Glaister, K. W., & Buckley, P. J. 1997. Task-related and partner-related selection criteria in UK international joint ventures.British Journal of Management, 8(3): 199–222.

Godfrey, P. C., & Hill, C. W. L. 1995. The problem of unobser-vables in strategic management research. Strategic Manage-ment Journal, 16(7): 519–533.

Gould, R. V. 2002. The origins of status hierarchies: A formaltheory and empirical test. American Journal of Sociology, 107(5):1143–1178.

Gulati, R., Lavie, D., & Singh, H. 2009. The nature of partneringexperience and the gains from alliances. Strategic ManagementJournal, 30(11): 1213–1233.

Gulati, R., & Westphal, J. D. 1999. Cooperative or controlling?The effects of CEO-board relations and the content of interlockson the formation of joint ventures. Administrative ScienceQuarterly, 44(3): 473–506.

Guler, I., & Guillén, M. F. 2010. Home country networks andforeign expansion: Evidence from the venture capital industry.Academy of Management Journal, 53(2): 390–410.

Heckman, J. J. 1979. Sample selection bias as a specification error.Econometrica, 47(1): 153–161.

Heil, O., & Robertson, T. S. 1991. Toward a theory of competitivemarket signaling: A research agenda. Strategic ManagementJournal, 12(6): 403–418.

Higgins, M. C., & Gulati, R. 2003. Getting off to a good start: Theeffects of upper echelon affiliations on underwriter prestige.Organization Science, 14(3): 244–263.

Hitt, M. A., Ahlstrom, D., Dacin, M. T., Levitas, E., & Svobodina, L.2004. The institutional effects on strategic alliance partnerselection in transition economies: China vs. Russia.OrganizationScience, 15(2): 173–185.

Isobe, T., Makino, S., & Montgemery, D. B. 2000. Resourcecommitment, entry timing, and market performance offoreign direct investments in emerging economies: The case ofJapanese international joint ventures in China. Academy ofManagement Journal, 43(3): 468–484.

Jacobson, R. 1992. The “Austrian” school of strategy. Academy ofManagement Review, 17(4): 782–807.

Jensen, M. 2008. The use of relational discrimination to managemarket entry: When do social status and structural holeswork against you? Academy of Management Journal, 51(4):723–743.

Jensen, M., & Roy, A. 2008. Staging exchange partner choices:When do status and reputation matter? Academy of Manage-ment Journal, 51(3): 495–516.

Keister, L. A. 2009. Organizational research on market transition:A sociological approach. Asia Pacific Journal of Management,26(4): 719–742.

Khoury, T., & Peng, M. W. 2011. Does institutional reform ofintellectual property rights lead to more inbound FDI? Evidencefrom Latin America and the Caribbean. Journal of World Busi-ness, 46(3): 337–345.

Kilduff, M., Crossland, C., Tsai, W., & Krackhardt, D. 2008.Organizational network perceptions vs reality: A small world



after all? Organizational Behavior and Human Decision Processes,107(1): 15–28.

Koka, B. R., & Prescott, J. E. 2008. Designing alliance networks:The influence of network position, environmental change, andstrategy on firm performance. Strategic Management Journal,29(6): 639–661.

Kostova, T., & Zaheer, S. 1999. Organizational legitimacy underconditions of complexity: The case of the multinational enter-prise. Academy of Management Review, 24(1): 64–81.

Krackhardt, D. 1987. Cognitive social structures. Social Networks,9(2): 109–134.

Krug, B., & Hendrischke, H. 2012. Market design in Chinesemarket places. Asia Pacific Journal of Management, 29(3):525–546.

Kwon, J.-W. 2012. Does China have more than one culture?Exploring regional differences of work value in China. AsiaPacific Journal of Management, 29(1): 79–102.

Laursen, K., Masciarelli, F., & Prencipe, A. 2012. Regionsmatter: How localized social capital affects innovation andexternal knowledge acquisition. Organization Science, 23(1):177–193.

Lavie, D. 2007. Alliance portfolios and firm performance: A studyof value creation and appropriation in the US software industry.Strategic Management Journal, 28(12): 1187–1212.

Levin, D. Z., & Cross, R. 2004. The strength of weak ties you cantrust: The mediating role of trust in effective knowledgetransfer. Management Science, 50(11): 1477–1490.

Li, H., & Zhang, Y. 2007. The role of managers’ political network-ing and functional experience in new venture performance:Evidence from China’s transition economy. Strategic Manage-ment Journal, 28(8): 791–804.

Li, J. J., Poppo, L., & Zhou, K. Z. 2008. Do managerial ties in Chinaalways produce value? Competition, uncertainty, and domesticvs. foreign firms. Strategic Management Journal, 29(4): 383–400.

Li, J., Zhou, C., & Zajac, E. J. 2009. Control, collaboration, andproductivity in international joint ventures: Theory and evi-dence. Strategic Management Journal, 30(8): 865–884.

Li, Y., Peng, M. W., & Macaulay, C. D. 2013. Market-politicalambidexterity during institutional transitions. Strategic Organi-zation, 11(2): 205–213.

Lin, Z., Peng, M. W., Yang, H., & Sun, S. L. 2009a. How donetworks and learning drive M&As? An institutional compar-ison of China and the United States. Strategic ManagementJournal, 30(10): 1113–1132.

Lin, Z., Yang, H., & Arya, B. 2009b. Alliance partners and firmperformance: Resource complementarity and status associa-tion. Strategic Management Journal, 30(9): 921–940.

Liu, X., Lu, J., Filatotchev, I., Buck, T., & Wright, M. 2010.Returnee entrepreneurs, knowledge spillovers and innovationin high-tech firms in emerging economies. Journal of Interna-tional Business Studies, 41(7): 1183–1197.

Liu, Y., Luo, Y., & Liu, T. 2009. Governing buyer–supplier rela-tionships through transactional and relational mechanisms:Evidence from China. Journal of Operations Management,27(4): 294–309.

Long, J. S., & Freese, J. 2006. Regression models for categorical andlimited dependent variables using stata, 2nd edn. College Sta-tion, TX: Stata Press.

Lu, J., & Ma, X. 2008. The contingent value of local partners’business group affiliations. Academy of Management Journal,51(2): 295–314.

Luo, X., & Chung, C. 2005. Keeping it all in the family: The role ofparticularistic relationships in business group performanceduring institutional transition. Administrative Science Quarterly,50(3): 404–439.

Luo, Y. 1997. Partner selection and venturing success: The case ofjoint ventures with firms in the People’s Republic of China.Organization Science, 8(6): 648–662.

Luo, Y. 1998. Joint venture success in China: How shouldwe select a good partner? Journal of World Business, 33(2):145–166.

Luo, Y. 2007. From foreign investors to strategic insiders: Shiftingparameters, prescriptions and paradigms for MNCs in China.Journal of World Business, 42(1): 14–34.

Luo, Y. 2008. Procedural fairness and interfirm cooperationin strategic alliances. Strategic Management Journal, 29(1):27–46.

Luo, Y., & Peng, M. W. 1999. Learning to compete in a transitioneconomy: Experience, environment, and performance. Journalof International Business Studies, 30(2): 269–295.

Luo, Y., & Tung, R. L. 2007. International expansion of emergingmarket enterprises: A springboard perspective. Journal of Inter-national Business Studies, 38(4): 481–498.

Madhavan, R., Gnyawali, D., & He, J. 2004. Two’s company,three’s a crowd? Triads in cooperative-competitive networks.Academy of Management Journal, 47(6): 918–927.

Mariotti, S., & Piscitello, L. 1995. Information costs and locationof FDIs within the host country: Empirical evidence from Italy.Journal of International Business Studies, 26(4): 815–841.

Markóczy, L., Sun, L., Peng, M. W., Shi, W., & Ren, B. 2013. Socialnetwork contingency, symbolic management, and boundarystretching. Strategic Management Journal, 34(11): 1367–1387.

Meyer, K. E., Estrin, S., Bhaumik, S. K., & Peng, M. W. 2009a.Institutions, resources, and entry strategies in emerging eco-nomies. Strategic Management Journal, 30(1): 61–80.

Meyer, K. E., & Nguyen, H. V. 2005. Foreign investment stra-tegies and sub-national institutions in emerging markets:Evidence from Vietnam. Journal of Management Studies, 42(1):63–93.

Meyer, K. E., & Sinani, E. 2009. When and where does foreigndirect investment generate positive spillovers: A meta-analysis.Journal of International Business Studies, 40(7): 1075–1094.

Meyer, K. E., Wright, M., & Pruthi, S. 2009b. Managing know-ledge in foreign entry strategies: A resource-based analysis.Strategic Management Journal, 30(5): 557–574.

Mitsuhashi, H., & Greve, H. 2009. A matching theory of allianceformation and organizational success: Complementarityand compatibility. Academy of Management Journal, 52(5):975–995.

Nair, A., Hanvanich, S., & Cavusgil, S. T. 2007. An exploration ofthe patterns underlying related and unrelated collaborativeventures using neural network: Empirical investigation of colla-borative venture formation data spanning 1985–2001. Interna-tional Business Review, 16(6): 659–686.

Nielsen, B. B. 2003. An empirical investigation of the drivers ofinternational strategic alliance formation. European Manage-ment Journal, 21(3): 301–322.

Obstfeld, D. 2005. Social networks, the tertius iungens orienta-tion, and involvement in innovation. Administrative ScienceQuarterly, 50(1): 100–130.

Okhmatovskiy, I. 2010. Performance implications of ties to thegovernment and SOEs: A political embeddedness perspective.Journal of Management Studies, 47(6): 1020–1047.

Owen-Smith, J., & Powell, W. W. 2004. Knowledge networksas channels and conduits: The effects of spillovers in theBoston biotechnology community. Organization Science,15(1): 5–21.

Ozcan, P., & Eisenhardt, K. M. 2009. Origin of alliance portfolios:Entrepreneurs, network strategies, and firm performance. Acad-emy of Management Journal, 52(2): 246–279.

Ozmel, U., Reuer, J. J., & Gulati, R. 2012. Signals across multiplenetworks: How venture capital and alliance networks affectinterorganizational collaboration. Academy of ManagementJournal, 56(3): 852–866.

Pajunen, K., & Fang, L. 2013. Dialectical tensions and pathdependence in international joint venture evolution and termi-nation. Asia Pacific Journal of Management, 30(2): 577–600.

Peng, M. W. 2003. Institutional transitions and strategic choices.Academy of Management Review, 28(2): 275–296.

Peng, M. W. 2004. Outside directors and firm performanceduring institutional transitions. Strategic Management Journal,25(5): 453–471.



Peng, M. W., Li, Y., Xie, E., & Su, Z. 2010. CEO duality,organizational slack, and firm performance in China. Asia PacificJournal of Management, 27(4): 611–624.

Peng, M. W., & Luo, Y. 2000. Managerial ties and firm perfor-mance in a transition economy: The nature of a micro–macrolink. Academy of Management Journal, 43(3): 486–501.

Podolny, J. M. 1993. A status-based model of market competi-tion. American Journal of Sociology, 98(4): 829–872.

Podolny, J. M. 2001. Networks as the pipes and prisms of themarket. American Journal of Sociology, 107(1): 33–60.

Polidoro, F. 2013. The competitive implications of certifications:The effects of scientific and regulatory certifications on entriesinto new technical fields. Academy of Management Journal,56(2): 597–627.

Pollock, T. G., & Gulati, R. 2007. Standing out from the crowd:The visibility-enhancing effects of IPO-related signals on alli-ance formation by entrepreneurial firms. Strategic Organization,5(4): 339–372.

Pollock, T. G., Rindova, V. P., & Maggitti, P. G. 2008. Marketwatch: Information and availability cascades among the mediaand investors in the US IPO market. Academy of ManagementJournal, 51(2): 335–358.

Puffer, S. M., McCarthy, D. J., & Peng, M. W. 2013. Managingfavors in a global economy. Asia Pacific Journal of Management,30(2): 321–326.

Ren, B., Au, K., & Birtch, T. 2009. China’s business networkstructure during institutional transitions. Asia Pacific Journal ofManagement, 26(2): 219–240.

Rindova, V. P., Williamson, I. O., Petkova, A. P., & Sever, J. M. 2005.Being good or being known: An empirical examination of thedimensions, antecedents, and consequences of organizationalreputation. Academy of Management Journal, 48(6): 1033–1049.

Roberts, J. H., & Lattin, J. M. 1997. Consideration: Review ofresearch and prospects for future insights. Journal of MarketingResearch, 34(3): 406–410.

Rowley, T., Behrens, D., & Krackhardt, D. 2000. Redundantgovernance structures: An analysis of structural and relationalembeddedness in the steel and semiconductor industries.Strategic Management Journal, 21(3): 369–386.

Roy, J.-P., & Oliver, C. 2009. International joint venture partnerselection: The role of the host-country legal environment.Journal of International Business Studies, 40(5): 779–801.

Shi, W., Markóczy, L., & Dess, G. G. 2009. The role of middle mana-gement in the strategy process: Group affiliation, structural holes,and tertius lungens. Journal of Management, 35(6): 1453–1480.

Shi, W., Sun, S. L., & Peng, M. W. 2012. Sub-national institutionalcontingencies, network positions, and IJV partner selection.Journal of Management Studies, 49(7): 1221–1245.

Shi, W., Markóczy, L., & Stan, C. 2013. The continuing importanceof political ties in China. Academy of Management Perspectives,published online May 24, doi: 10.5465/amp.2011.0153.

Shipilov, A. V., & Li, S. X. 2008. Can you have your cake and eat ittoo? Structural holes’ influence on status accumulation andmarket performance in collaborative networks. AdministrativeScience Quarterly, 53(1): 73–108.

Siegel, J. 2007. Contingent political capital and internationalalliances: Evidence from South Korea. Administrative ScienceQuarterly, 52(4): 621–666.

Spicer, A., Dunfee, T. W., & Bailey, W. J. 2004. Does nationalcontext matter in ethical decision making? An empirical test ofintegrative social contracts theory. Academy of ManagementJournal, 47(4): 610–620.

Stan, C., Peng, M. W., & Bruton, G. D. 2013. Slack andthe performance of state-owned enterprises. Asia Pacific Journalof Management, published online 27 March, doi:10.1007/s10490-013-9347-7.

Stiglitz, J. E. 2000. The contributions of the economics ofinformation to twentieth century economics. The QuarterlyJournal of Economics, 115(4): 1441–1478.

Sun, S. L., Chen, H., & Pleggenkuhle-Miles, E. 2010. Movingupward in global value chains: The innovation of mobile

phone developers in China. Chinese Management Studies, 4(4):305–321.

Sun, S. L., & Lee, R. P. 2013. Enhancing innovation throughinternational joint venture portfolios: From the emerging firmperspective. Journal of International Marketing, 21(3): 1–21.

Sun, S. L., Peng, M. W., Ren, B., & Yan, D. 2012. A comparativeownership advantage framework for cross-border M&As: Therise of Chinese and Indian MNEs. Journal of World Business,47(1): 4–16.

Sytch, M., Tatarynowicz, A., & Gulati, R. 2012. Toward a theory ofextended contract: The incentives and opportunities for brid-ging across network communities. Organization Science, 23(6):1651–1658.

Tan, D., & Meyer, K. E. 2011. Country-of-origin an industry FDIagglomeration of foreign investors in an emerging economy.Journal of International Business Studies, 42(4): 504–520.

Tatoglu, E., & Glaister, K. W. 2000. Strategic motives and partnerselection criteria in international joint ventures in Turkey:Perspectives of Western firms and Turkish firms. Journal of GlobalMarketing, 13(3): 53–92.

Teagarden, M. B., & Schotter, A. 2013. Favor prevalence inemerging markets: A multi-level analysis. Asia Pacific Journal ofManagement, 30(2): 447–460.

Thams, Y., Liu, Y., & Von Glinow, M. A. 2013. Asian favors: Morethan a cookie cutter approach. Asia Pacific Journal of Manage-ment, 30(2): 461–486.

Tong, T. W., Reuer, J. J., & Peng, M. W. 2008. International jointventures and the value of growth options. Academy of Manage-ment Journal, 51(5): 1014–1029.

Tracey, P., & Phillips, N. 2011. Entrepreneurship in emergingmarkets: Strategies for new venture creation in uncertaininstitutional contexts. Management International Review, 51(1):23–39.

Uzzi, B., & Lancaster, R. 2005. Relational embeddedness andlearning: The case of bank loan managers and their clients.Management Science, 49(4): 383–399.

von Hecker, U. 1993. The importance of balance, generalizationand positivity as cognitive structural rules. Zeitschrift fur Experi-mentelle und Angewandte Psychologie, 40(1): 548–576.

Wasserman, S., & Faust, K. 1994. Social network analysis: Methodsand applications. Cambridge, UK: Cambridge University Press.

Winter, S. G. 2012. Purpose and progress in the theory of strategy:Comments on Gavetti. Organization Science, 23(1): 288–297.

Wolf, R. C. 2000. Effective international joint venture management:Practical legal insights for successful organization and implemen-tation. Armonk, NY: M.E. Sharpe.

Wong, L. K., & Ellis, P. 2002. Social ties and partner identificationin Sino–Hong Kong international joint ventures. Journal ofInternational Business Studies, 33(2): 267–289.

Wooldridge, J. M. 2006. Introductory econometrics: A modernapproach. Mason, OH: South-Western.

Xia, J., Tan, J., & Tan, D. 2008. Mimetic entry and bandwagoneffect: The rise and decline of international equity joint venturein China. Strategic Management Journal, 29(2): 195–217.

Xiao, Z. X., & Tsui, A. S. 2007. When brokers may not work: Thecultural contingency of social capital in Chinese high-techfirms. Administrative Science Quarterly, 52(1): 1–31.

Xu, D., & Shenkar, O. 2002. Institutional distance and themultinational enterprise. Academy of Management Review,27(4): 608–618.

Yan, R. 1998. Short-term results: The litmus test for success inChina. Harvard Business Review, 76(5): 61–69.

Yang, H., Lin, Z., & Peng, M. W. 2011a. Behind acquisitions ofalliance partners: Exploratory learning and network embedd-edness. Academy of Management Journal, 54(5): 1069–1080.

Yang, H., Sun, S. L., Lin, Z., & Peng, M. W. 2011b. BehindM&As in China and the United States: Networks, learning,and institutions. Asia Pacific Journal of Management, 28(2):239–255.

Zaheer, S. 1995. Overcoming the liability of foreignness. Academyof Management Journal, 38(2): 341–363.



http://dx.doi.org/10.1007/s10490-013-9347-7).

http://dx.doi.org/10.1007/s10490-013-9347-7).

Zavyalova, A., Pfarrer, M. D., Reger, R. K., & Shapiro, D. L. 2012.Managing the message: The effects of firm actions and industryspillovers on media coverage following wrongdoing. Academyof Management Journal, 55(5): 1079–1101.

Zhan, W., & Chen, R. 2013. Dynamic capability and IJV perfor-mance: The effect of exploitation and exploration capabilities.Asia Pacific Journal of Management, 30(2): 601–632.

Zhan, W., Chen, R., Erramilli, M. K., & Nguyen, D. T. 2009.Acquisition of organizational capabilities and competitiveadvantage of IJVs in transition economies: The case of Vietnam.Asia Pacific Journal of Management, 26(2): 285–308.

Zhao, Z., Anand, J., & Mitchell, W. 2005. A dual networksperspective on inter-organizational transfer of R&D capabilities:International joint ventures in the Chinese automotive industry.Journal of Management Studies, 42(1): 127–160.

ABOUT THE AUTHORSWeilei (Stone) Shi (PhD, University of Pittsburgh) isan Assistant Professor of Strategy at Zicklin School ofBusiness, Baruch College, City University of NewYork. His research interests center on the interactionbetween strategy and international management.Specifically, Stone is interested in examining M&Aand alliances from both temporal and network per-spectives. Before joining academia Stone worked forRoland Berger Strategy Consultancy, a global strategyconsulting firm headquartered in Munich, Germany.

Sunny Li Sun (PhD, University of Texas at Dallas) isan Assistant Professor of Entrepreneurship and Inno-vation at the University of Missouri, Kansas City’sHenry W. Bloch School of Management. Hisresearch interests include new venture creation andfinancing, business models, entrepreneurial firmgrowth, internationalization, and corporate govern-ance. He is a Chinese citizen.

Brian C Pinkham (PhD, University of Texas atDallas) is an Assistant Professor of InternationalBusiness at the Richard Ivey Business School, Uni-versity of Western Ontario. His research interestare global strategy, international business, and legalinstitutions.

MikeW Peng (PhD, University ofWashington) is theJindal Chair of Global Strategy at the Jindal School ofManagement, University of Texas at Dallas; aNational Science Foundation CAREER Awardwinner; and a Fellow of the Academy of InternationalBusiness. His research interests are global strategy,international business, and emerging economies,with a focus on the institution-based view.

Accepted by Yadong Luo, Consulting Editor, 5 December 2013. This article has been with the authors for four revisions.



domestic alliance network to attract foreign partners: evidence from

Documents